Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Utility to SMF a new demon/service ?
94 views
Skip to first unread message
Philip Brown
Mar 21, 2012, 2:16:56 PM3/21/12
to
I'm not a big redhat fan, but I bring up something they do as a point of reference.
Turning (random process) into a boot-time service, is trivial.
1. Make a script for it in /etc/init.d that supports "start" and "stop" options.
2, run "chkconfig (newname) on"
DONE.
On solaris, it is painful. You still have part 1, make an init script... but then the docs tell you to wade through a mountain of xml. Yuck!
Is there no common "make an SMF service" helper out there yet? My searches did not yield fruit.
It would be really surprising if there still isnt one yet, after 7 years of SMF
Turning (random process) into a boot-time service, is trivial.
1. Make a script for it in /etc/init.d that supports "start" and "stop" options.
2, run "chkconfig (newname) on"
DONE.
On solaris, it is painful. You still have part 1, make an init script... but then the docs tell you to wade through a mountain of xml. Yuck!
Is there no common "make an SMF service" helper out there yet? My searches did not yield fruit.
It would be really surprising if there still isnt one yet, after 7 years of SMF
cindy
Mar 21, 2012, 3:26:12 PM3/21/12
to
Oscar del Rio
Mar 21, 2012, 3:53:43 PM3/21/12
to
On 03/21/12 02:16 PM, Philip Brown wrote:
> Is there no common "make an SMF service" helper out there yet?
have you tried "manifold"?
> Is there no common "make an SMF service" helper out there yet?
http://code.google.com/p/manifold/
Oscar del Rio
Mar 21, 2012, 4:09:28 PM3/21/12
to
On 03/21/12 02:16 PM, Philip Brown wrote:
cp myscript /etc/init.d
ln /etc/init.d/myscrypt /etc/rc3.d/S99myscript
/etc/init.d/myscript start
It should work on any version of Solaris; you just don't get the
benefits of SMF
Philip Brown
Mar 21, 2012, 5:40:13 PM3/21/12
to
Philip Brown
Mar 21, 2012, 5:39:41 PM3/21/12
to
On Wednesday, March 21, 2012 12:26:12 PM UTC-7, cindy wrote:
> On Mar 21, 12:16 pm, Philip Brown
> On Mar 21, 12:16 pm, Philip Brown
I specifically asked for a 'helper' (app).
Happily, Oscar provided a reference to one.
Casper H.S. Dik
Mar 22, 2012, 4:01:50 AM3/22/12
to
service XML file and change the name, dependencies, etc.
Import them (svcadm restart manifest-import)
(Clearly, this should be made simpler)
Casper
Philip Brown
Mar 22, 2012, 12:25:15 PM3/22/12
to
On Thursday, March 22, 2012 1:01:50 AM UTC-7, Casper H. S. Dik wrote:
> Philip Brown
"It's not that hard to customise this program. Extract the source code, change the #defines in the right places, and type 'make'"
That also falls in the realm of "not that difficult".
But it's not a very good 'customer interface' either.
> (Clearly, this should be made simpler)
I think this is the important take-away from this :) It's long past time that sun^H^H^Horacle provided a decent end-user tool.
"Edit the xml file" is NOT "a decent end-user tool".
Especially one as convoluted as an SMF manifest.
Take a look at the user-level questions that
http://code.google.com/p/manifold/
asks when you run it.
Manifests could have an initial section of a few lines that hold those values, at the top. Rather than having to go through the entire file, "parsing" the xml in your head.
If xml is sooo simple to parse, why is libxml2.so 2 megabytes! :P
It's not nice to force humans to do something, that takes a machine many many lines of code to do.
ugh.
Personally, I would prefer a tool even simpler than the "manifold" tool. Something as simple as the redhat "chkconfig" tool.
That is to say, default to handling the 95% of most sysadmin's needs, with a minimum of unneccessary questions.
It's almost always just a simple case of, "hey, I want this demon started at boot time, here's the start/stop script".
Oracle is not meeting customer needs well in this area.
> Philip Brown
>
> >Is there no common "make an SMF service" helper out there yet? My searches did not yield fruit.
> >It would be really surprising if there still isnt one yet, after 7 years of SMF
>
> I've done it a few time and it isn't all that difficult: copy a similar
> service XML file and change the name, dependencies, etc.
> Import them (svcadm restart manifest-import)
that is semantically almost identical to,
> >Is there no common "make an SMF service" helper out there yet? My searches did not yield fruit.
> >It would be really surprising if there still isnt one yet, after 7 years of SMF
>
> I've done it a few time and it isn't all that difficult: copy a similar
> service XML file and change the name, dependencies, etc.
> Import them (svcadm restart manifest-import)
"It's not that hard to customise this program. Extract the source code, change the #defines in the right places, and type 'make'"
That also falls in the realm of "not that difficult".
But it's not a very good 'customer interface' either.
> (Clearly, this should be made simpler)
"Edit the xml file" is NOT "a decent end-user tool".
Especially one as convoluted as an SMF manifest.
Take a look at the user-level questions that
http://code.google.com/p/manifold/
asks when you run it.
Manifests could have an initial section of a few lines that hold those values, at the top. Rather than having to go through the entire file, "parsing" the xml in your head.
If xml is sooo simple to parse, why is libxml2.so 2 megabytes! :P
It's not nice to force humans to do something, that takes a machine many many lines of code to do.
ugh.
Personally, I would prefer a tool even simpler than the "manifold" tool. Something as simple as the redhat "chkconfig" tool.
That is to say, default to handling the 95% of most sysadmin's needs, with a minimum of unneccessary questions.
It's almost always just a simple case of, "hey, I want this demon started at boot time, here's the start/stop script".
Oracle is not meeting customer needs well in this area.
Philip Brown
Mar 22, 2012, 12:26:43 PM3/22/12
to
On Thursday, March 22, 2012 9:25:15 AM UTC-7, Philip Brown wrote:
>
> It's almost always just a simple case of, "hey, I want this demon started at boot time, here's the start/stop script".
>
> Oracle is not meeting customer needs well in this area.
PS: no, "hey you can still use /etc/rc*/S*" is not a good reply to this.
>
> It's almost always just a simple case of, "hey, I want this demon started at boot time, here's the start/stop script".
>
> Oracle is not meeting customer needs well in this area.
Oracle is supposedly trying to get rid of those things, so telling customers "go use that instead" is not appropriate.
Make the "new technology" more customer friendly please.
Chris Ridd
Mar 22, 2012, 1:31:01 PM3/22/12
to
On 2012-03-22 16:25:15 +0000, Philip Brown said:
> It's almost always just a simple case of, "hey, I want this demon
> started at boot time, here's the start/stop script".
>
> Oracle is not meeting customer needs well in this area.
manifold has been around for several years - it is surprising it never
> It's almost always just a simple case of, "hey, I want this demon
> started at boot time, here's the start/stop script".
>
> Oracle is not meeting customer needs well in this area.
got considered for incorporation into OpenSolaris.
--
Chris
Ian Collins
Mar 22, 2012, 2:13:24 PM3/22/12
to
On 03/23/12 05:25 AM, Philip Brown wrote:
>
> Personally, I would prefer a tool even simpler than the "manifold" tool. Something as simple as the redhat "chkconfig" tool.
> That is to say, default to handling the 95% of most sysadmin's needs, with a minimum of unneccessary questions.
>
> It's almost always just a simple case of, "hey, I want this demon started at boot time, here's the start/stop script".
>
> Oracle is not meeting customer needs well in this area.
I guess most people who add services are developers, so a quick edit of
>
> Personally, I would prefer a tool even simpler than the "manifold" tool. Something as simple as the redhat "chkconfig" tool.
> That is to say, default to handling the 95% of most sysadmin's needs, with a minimum of unneccessary questions.
>
> It's almost always just a simple case of, "hey, I want this demon started at boot time, here's the start/stop script".
>
> Oracle is not meeting customer needs well in this area.
an XML file isn't daunting. It is just another form of configuration
file, isn't it?
--
Ian Collins
Philip Brown
Mar 22, 2012, 5:44:54 PM3/22/12
to
I have a CS degree.
I have written quite a few programs. including .. what.. 10? solaris device drivers now?
I still dont WANT to be in "developer mode", to do something that should be trivial, like add a service.
As I've pointed out: redhat makes it easy. They recognize the opposite: they come from the assumption that most people who are adding services, are NOT developers, and they give them a one-line method of adding a service.
(one line, if they are dropping in a 3rd party "init" script from a 3rd party app. which is fairly common)
btw, for the record, no, I dont consider XML to be "just another form of configuration file", just as I dont consider a .emacs file of any decent length, to be "just another form of configuration file".
Both are more akin to programming, than mere config files.
There is a very good reason why *most* programs that decide to have their config files in xml format, also provide a user-friendly GUI, so that normal people dont have to hand-edit XML code.
One might say that zone configuration is, in its own way, LESS complex, than SMF manifests.
The average zone xml config file is a mere 20 lines, compared with 50-100 lines of an average manifest file.
Yet solaris provides 'zonecfg', rather than telling people "oh hey, just go create yourzone.xml under /etc/zones -- here's a template for you to start from".
As I have said: not meeting customer needs here.
Ian Collins
Mar 22, 2012, 11:36:28 PM3/22/12
to
On 03/23/12 10:44 AM, Philip Brown wrote:
> On Thursday, March 22, 2012 11:13:24 AM UTC-7, Ian Collins wrote:
>> On 03/23/12 05:25 AM, Philip Brown wrote:
>>
>>> It's almost always just a simple case of, "hey, I want this demon started at boot time, here's the start/stop script".
>>>
>>> Oracle is not meeting customer needs well in this area.
>>
>> I guess most people who add services are developers, so a quick edit of
>> an XML file isn't daunting. It is just another form of configuration
>> file, isn't it?
>>
>
> I can be considered "a developer".
> I have a CS degree.
> I have written quite a few programs. including .. what.. 10? solaris device drivers now?
Which probably have config files...
> On Thursday, March 22, 2012 11:13:24 AM UTC-7, Ian Collins wrote:
>> On 03/23/12 05:25 AM, Philip Brown wrote:
>>
>>> It's almost always just a simple case of, "hey, I want this demon started at boot time, here's the start/stop script".
>>>
>>> Oracle is not meeting customer needs well in this area.
>>
>> I guess most people who add services are developers, so a quick edit of
>> an XML file isn't daunting. It is just another form of configuration
>> file, isn't it?
>>
>
> I can be considered "a developer".
> I have a CS degree.
> I have written quite a few programs. including .. what.. 10? solaris device drivers now?
> I still dont WANT to be in "developer mode", to do something that should be trivial, like add a service.
> As I've pointed out: redhat makes it easy. They recognize the opposite: they come from the assumption that most people who are adding services, are NOT developers, and they give them a one-line method of adding a service.
> (one line, if they are dropping in a 3rd party "init" script from a 3rd party app. which is fairly common)
> btw, for the record, no, I dont consider XML to be "just another form of configuration file", just as I dont consider a .emacs file of any decent length, to be "just another form of configuration file".
> Both are more akin to programming, than mere config files.
I spend a lot of time with XML, so our opinions on that are bound to differ.
> One might say that zone configuration is, in its own way, LESS complex, than SMF manifests.
> The average zone xml config file is a mere 20 lines, compared with 50-100 lines of an average manifest file.
>
> Yet solaris provides 'zonecfg', rather than telling people "oh hey, just go create yourzone.xml under /etc/zones -- here's a template for you to start from".
than adding services. Although I must admit I have a replication
service I often add to new machines or zones and I have written a simple
tool to generate host specific manifests.
> As I have said: not meeting customer needs here.
--
Ian Collins
Philip Brown
Mar 23, 2012, 2:40:54 PM3/23/12
to
On Thursday, March 22, 2012 8:36:28 PM UTC-7, Ian Collins wrote:
> On 03/23/12 10:44 AM, Philip Brown wrote:
>
> > Yet solaris provides 'zonecfg', rather than telling people "oh hey, just go create yourzone.xml under /etc/zones -- here's a template for you to start from".
> ...
> On 03/23/12 10:44 AM, Philip Brown wrote:
>
> > Yet solaris provides 'zonecfg', rather than telling people "oh hey, just go create yourzone.xml under /etc/zones -- here's a template for you to start from".
>
> > As I have said: not meeting customer needs here.
>
> Maybe (paying) customers simply haven't asked?
>
Oh come on now... are you seriously suggesting, that in **5 years** since it came out, NO-ONE with a support contract, has "asked" for this to be improved?
> > As I have said: not meeting customer needs here.
>
> Maybe (paying) customers simply haven't asked?
>
And that "all we have to do is ask"?
If you are really claiming this, please make a public statement here as to how many support-contract-backed requests it will take, to see something come out of oracle in a reasonable amount of time, and I think we the paying customers will make it clear that we would like to see it.
Ian Collins
Mar 23, 2012, 3:49:46 PM3/23/12
to
--
Ian Collins
Richard B. Gilbert
Mar 23, 2012, 4:16:48 PM3/23/12
to
attention to your needs. If not, they don't owe you anything! That
doesn't mean that they won't do anything but it's unlikely to get
anything done in a short time.
Richard B. Gilbert
Mar 23, 2012, 5:00:03 PM3/23/12
to
Do you have a support contract?
Have you asked for this enhancement and explained why you think it is
needed? I suspect that Oracle/Sun would at least consider doing it if
enough paying customers requested it.
This is not really the proper forum for your request. Try requesting
it from someone who can DO something about it. A few Sun people
participate in this newsgroup but they are, for the most part, technical
people. The decision makers do not, for the most part, read this
newsgroup! Those who do might not be able to do anything more than
passing it to somebody who CAN do something.
John D Groenveld
Mar 23, 2012, 9:19:57 PM3/23/12
to
In article <23438456.671.1332528054357.JavaMail.geo-discussion-forums@ynjx8>,
it here so other customers can add their CSI numbers to it.
John
groe...@acm.org
Philip Brown <ph...@bolthole.com> wrote:
>Oh come on now... are you seriously suggesting, that in **5 years**
>since it came out, NO-ONE with a support contract, has "asked" for this
>to be improved?
If you get a CR or RFE from Chuck Rozwat and company, please post
>Oh come on now... are you seriously suggesting, that in **5 years**
>since it came out, NO-ONE with a support contract, has "asked" for this
>to be improved?
it here so other customers can add their CSI numbers to it.
John
groe...@acm.org
Cydrome Leader
Mar 24, 2012, 10:33:35 PM3/24/12
to
good job oracle.
Cydrome Leader
Mar 24, 2012, 10:35:01 PM3/24/12
to
Ian Collins
Mar 24, 2012, 10:40:24 PM3/24/12
to
--
Ian Collins
Cydrome Leader
Mar 24, 2012, 10:43:48 PM3/24/12
to
I opened a case about a bug where entering NTP servers into a jumpstart config causes the install
to go interactive and prompt for a date and time.
bug 7059983 has been open since June 21 of last year. Still no fix. I do get updates every two
weeks or so. Development is working on it now though.
YTC#1
Mar 26, 2012, 7:41:22 AM3/26/12
to
On 03/23/12 03:36 AM, Ian Collins wrote:
> On 03/23/12 10:44 AM, Philip Brown wrote:
>> On Thursday, March 22, 2012 11:13:24 AM UTC-7, Ian Collins wrote:
>>> On 03/23/12 05:25 AM, Philip Brown wrote:
>>>
<snip>
> On 03/23/12 10:44 AM, Philip Brown wrote:
>> On Thursday, March 22, 2012 11:13:24 AM UTC-7, Ian Collins wrote:
>>> On 03/23/12 05:25 AM, Philip Brown wrote:
>>>
>> As I have said: not meeting customer needs here.
>
> Maybe (paying) customers simply haven't asked?
>
Since day 1
--
Bruce Porter
XJR1300SP, XJ900F, Pegaso 650 Trail (x2) one red one grey
POTM#1(KoTL), WUSS#1 , YTC#1(bar), OSOS#2(KoTL) , DS#3 , IbW#18 ,Apostle#8
"The internet is a huge and diverse community but mainly friendly"
http://blog.maui.co.uk/index.php/ytc/
There *is* an alternative! http://www.openoffice.org/
YTC#1
Mar 26, 2012, 7:43:19 AM3/26/12
to
On 03/22/12 06:13 PM, Ian Collins wrote:
> On 03/23/12 05:25 AM, Philip Brown wrote:
>>
>> Personally, I would prefer a tool even simpler than the "manifold"
>> tool. Something as simple as the redhat "chkconfig" tool.
>> That is to say, default to handling the 95% of most sysadmin's needs,
>> with a minimum of unneccessary questions.
>>
>> It's almost always just a simple case of, "hey, I want this demon
>> started at boot time, here's the start/stop script".
>>
>> Oracle is not meeting customer needs well in this area.
>
> I guess most people who add services are developers, so a quick edit of
Not really, admins do it.
> On 03/23/12 05:25 AM, Philip Brown wrote:
>>
>> Personally, I would prefer a tool even simpler than the "manifold"
>> tool. Something as simple as the redhat "chkconfig" tool.
>> That is to say, default to handling the 95% of most sysadmin's needs,
>> with a minimum of unneccessary questions.
>>
>> It's almost always just a simple case of, "hey, I want this demon
>> started at boot time, here's the start/stop script".
>>
>> Oracle is not meeting customer needs well in this area.
>
> I guess most people who add services are developers, so a quick edit of
> an XML file isn't daunting. It is just another form of configuration
> file, isn't it?
>
No it isn't
/etc/hostname.<nic> is a config file
That was a nice easy thing to use in SunOS until S11
--
Bruce Porter
Rich Teer
Mar 26, 2012, 11:47:10 AM3/26/12
to
On Mon, 26 Mar 2012, YTC#1 wrote:
> /etc/hostname.<nic> is a config file
>
> That was a nice easy thing to use in SunOS until S11
I must admit (and it seems I'm not alone here), but as Solaris
> /etc/hostname.<nic> is a config file
>
> That was a nice easy thing to use in SunOS until S11
moves further and further away from plain old text config files
(in the grand old UNIX tradition), us old timers appreciate it
less and less. What next, a registry a la Windoze?!
Give me plain text files and vi over n (for suitably large values
of n) bespoke tools for editing opaque XML (or other) files!
--
Rich Teer, Publisher
Vinylphile Magazine
www.vinylphilemag.com
YTC#1
Mar 26, 2012, 12:52:23 PM3/26/12
to
On 03/26/12 04:47 PM, Rich Teer wrote:
> On Mon, 26 Mar 2012, YTC#1 wrote:
>
>> /etc/hostname.<nic> is a config file
>>
>> That was a nice easy thing to use in SunOS until S11
>
> I must admit (and it seems I'm not alone here), but as Solaris
> moves further and further away from plain old text config files
> (in the grand old UNIX tradition), us old timers appreciate it
I don't class myself as old, yet :-)
> On Mon, 26 Mar 2012, YTC#1 wrote:
>
>> /etc/hostname.<nic> is a config file
>>
>> That was a nice easy thing to use in SunOS until S11
>
> I must admit (and it seems I'm not alone here), but as Solaris
> moves further and further away from plain old text config files
> (in the grand old UNIX tradition), us old timers appreciate it
> less and less. What next, a registry a la Windoze?!
>
> Give me plain text files and vi over n (for suitably large values
> of n) bespoke tools for editing opaque XML (or other) files!
>
And lets not have servers install with NWAM/DHCP by default as well.
--
Bruce Porter
Cydrome Leader
Mar 26, 2012, 1:05:21 PM3/26/12
to
stupid fuckers.
YTC#1
Mar 26, 2012, 1:48:47 PM3/26/12
to
A combination of netadm and ipadm are your new friends
>
> stupid fuckers.
I don't think the people responsible are stupid.
Cydrome Leader
Mar 26, 2012, 2:02:55 PM3/26/12
to
Rich Teer <rich...@rite-group.com> wrote:
> On Mon, 26 Mar 2012, YTC#1 wrote:
>
>> /etc/hostname.<nic> is a config file
>>
>> That was a nice easy thing to use in SunOS until S11
>
> I must admit (and it seems I'm not alone here), but as Solaris
> moves further and further away from plain old text config files
> (in the grand old UNIX tradition), us old timers appreciate it
> less and less. What next, a registry a la Windoze?!
>
> Give me plain text files and vi over n (for suitably large values
> of n) bespoke tools for editing opaque XML (or other) files!
or at least netinfo, like next had.
> On Mon, 26 Mar 2012, YTC#1 wrote:
>
>> /etc/hostname.<nic> is a config file
>>
>> That was a nice easy thing to use in SunOS until S11
>
> I must admit (and it seems I'm not alone here), but as Solaris
> moves further and further away from plain old text config files
> (in the grand old UNIX tradition), us old timers appreciate it
> less and less. What next, a registry a la Windoze?!
>
> Give me plain text files and vi over n (for suitably large values
> of n) bespoke tools for editing opaque XML (or other) files!
you edit text files, then commit them back into the DB.
John D Groenveld
Mar 26, 2012, 4:26:33 PM3/26/12
to
In article <jkq6tg$dil$1...@dont-email.me>, YTC#1 <b...@ytc1-spambin.co.uk> wrote:
>And lets not have servers install with NWAM/DHCP by default as well.
Banged my head against the DHCP (NWAM) selection in the text
>And lets not have servers install with NWAM/DHCP by default as well.
installer.
NWAM breaks ipadm(1M):
<URL:http://docs.oracle.com/cd/E23824_01/html/E23811/glmdq.html#glpbq>
Shame on me for the reading the fucking manual but
Oracle doesn't sell mobile workstation systems and
doesn't seem much interested in supporting a vibrant
third-party Solaris developer ecosystem, so NWAM seems
superfluous.
John
groe...@acm.org
Ian Collins
Mar 27, 2012, 12:17:09 AM3/27/12
to
On 03/27/12 12:41 AM, YTC#1 wrote:
> On 03/23/12 03:36 AM, Ian Collins wrote:
>> On 03/23/12 10:44 AM, Philip Brown wrote:
>>> On Thursday, March 22, 2012 11:13:24 AM UTC-7, Ian Collins wrote:
>>>> On 03/23/12 05:25 AM, Philip Brown wrote:
>>>>
> <snip>
>
>>> As I have said: not meeting customer needs here.
>>
>> Maybe (paying) customers simply haven't asked?
>>
>
> They have
Care to be more specific?
> On 03/23/12 03:36 AM, Ian Collins wrote:
>> On 03/23/12 10:44 AM, Philip Brown wrote:
>>> On Thursday, March 22, 2012 11:13:24 AM UTC-7, Ian Collins wrote:
>>>> On 03/23/12 05:25 AM, Philip Brown wrote:
>>>>
> <snip>
>
>>> As I have said: not meeting customer needs here.
>>
>> Maybe (paying) customers simply haven't asked?
>>
>
> They have
--
Ian Collins
Cydrome Leader
Mar 27, 2012, 12:52:33 AM3/27/12
to
YTC#1 <b...@ytc1-spambin.co.uk> wrote:
> On 03/26/12 06:05 PM, Cydrome Leader wrote:
>> YTC#1<b...@ytc1-spambin.co.uk> wrote:
>>> On 03/22/12 06:13 PM, Ian Collins wrote:
>>>> On 03/23/12 05:25 AM, Philip Brown wrote:
>>>>>
>>>>> Personally, I would prefer a tool even simpler than the "manifold"
>>>>> tool. Something as simple as the redhat "chkconfig" tool.
>>>>> That is to say, default to handling the 95% of most sysadmin's needs,
>>>>> with a minimum of unneccessary questions.
>>>>>
>>>>> It's almost always just a simple case of, "hey, I want this demon
>>>>> started at boot time, here's the start/stop script".
>>>>>
>>>>> Oracle is not meeting customer needs well in this area.
>>>>
>>>> I guess most people who add services are developers, so a quick edit of
>>>
>>> Not really, admins do it.
>>>
>>>> an XML file isn't daunting. It is just another form of configuration
>>>
>>> Yes it is.
>>>
>>>> file, isn't it?
>>>>
>>>
>>> No it isn't
>>>
>>> /etc/hostname.<nic> is a config file
>>>
>>> That was a nice easy thing to use in SunOS until S11
>>
>> did they xml that too now?
>
> It doesn't exist any more.
>
> A combination of netadm and ipadm are your new friends
Cute new commands, but they should be fixing my NTP jumpstart bug first,
> On 03/26/12 06:05 PM, Cydrome Leader wrote:
>> YTC#1<b...@ytc1-spambin.co.uk> wrote:
>>> On 03/22/12 06:13 PM, Ian Collins wrote:
>>>> On 03/23/12 05:25 AM, Philip Brown wrote:
>>>>>
>>>>> Personally, I would prefer a tool even simpler than the "manifold"
>>>>> tool. Something as simple as the redhat "chkconfig" tool.
>>>>> That is to say, default to handling the 95% of most sysadmin's needs,
>>>>> with a minimum of unneccessary questions.
>>>>>
>>>>> It's almost always just a simple case of, "hey, I want this demon
>>>>> started at boot time, here's the start/stop script".
>>>>>
>>>>> Oracle is not meeting customer needs well in this area.
>>>>
>>>> I guess most people who add services are developers, so a quick edit of
>>>
>>> Not really, admins do it.
>>>
>>>> an XML file isn't daunting. It is just another form of configuration
>>>
>>> Yes it is.
>>>
>>>> file, isn't it?
>>>>
>>>
>>> No it isn't
>>>
>>> /etc/hostname.<nic> is a config file
>>>
>>> That was a nice easy thing to use in SunOS until S11
>>
>> did they xml that too now?
>
> It doesn't exist any more.
>
> A combination of netadm and ipadm are your new friends
or updating the 1980s style routing deficiencies, or vi bug from 1984 as
somebody here determined.
YTC#1
Mar 27, 2012, 5:56:24 AM3/27/12
to
have upset a lot of people.
From the workstation/laptop perspective I'd like to hear from Casper,
as he has always been an advocate of using Solaris on a laptop.
I still live by Solaris (s11) on my laptop and Desktop, sometimes
painful, but only when I have a config type issue.
I suspect drivers are going to be harder to come by.
YTC#1
Mar 27, 2012, 5:59:20 AM3/27/12
to
towards SMF using XML was apparent a lot of people moaned and asked for
it to be made easier.
We usually just pointed them to the sales/support people. From a PS
perspective SMF was a (small) money earner.
A PITA though, and I can never remember what bits go where without
looking at an already working one.
--
Bruce Porter
Casper H.S. Dik
Mar 27, 2012, 7:41:22 AM3/27/12
to
YTC#1 <b...@ytc1-spambin.co.uk> writes:
>On 03/26/12 09:26 PM, John D Groenveld wrote:
>> In article<jkq6tg$dil$1...@dont-email.me>, YTC#1<b...@ytc1-spambin.co.uk> wrote:
>>> And lets not have servers install with NWAM/DHCP by default as well.
>>
>> Banged my head against the DHCP (NWAM) selection in the text
>> installer.
>> NWAM breaks ipadm(1M):
>> <URL:http://docs.oracle.com/cd/E23824_01/html/E23811/glmdq.html#glpbq>
>>
>> Shame on me for the reading the fucking manual but
>> Oracle doesn't sell mobile workstation systems and
>> doesn't seem much interested in supporting a vibrant
>> third-party Solaris developer ecosystem, so NWAM seems
>> superfluous.
>I don't think I have ever seen so much swearing in this NG. S11 seems to
>have upset a lot of people.
> From the workstation/laptop perspective I'd like to hear from Casper,
>as he has always been an advocate of using Solaris on a laptop.
I'm still using solaris on my laptop.
>On 03/26/12 09:26 PM, John D Groenveld wrote:
>> In article<jkq6tg$dil$1...@dont-email.me>, YTC#1<b...@ytc1-spambin.co.uk> wrote:
>>> And lets not have servers install with NWAM/DHCP by default as well.
>>
>> Banged my head against the DHCP (NWAM) selection in the text
>> installer.
>> NWAM breaks ipadm(1M):
>> <URL:http://docs.oracle.com/cd/E23824_01/html/E23811/glmdq.html#glpbq>
>>
>> Shame on me for the reading the fucking manual but
>> Oracle doesn't sell mobile workstation systems and
>> doesn't seem much interested in supporting a vibrant
>> third-party Solaris developer ecosystem, so NWAM seems
>> superfluous.
>I don't think I have ever seen so much swearing in this NG. S11 seems to
>have upset a lot of people.
> From the workstation/laptop perspective I'd like to hear from Casper,
>as he has always been an advocate of using Solaris on a laptop.
In Solaris 10 it was even more difficult, especially when you have ethernet
and wifi. Before that, I needed to had my own "find my location" software
which barely worked.
Unfortunately, at the end of the Solaris 11 release some changes were
made which weren't perhaps complete. While I was able to get the changes to
setting the zone "fixed" ("zic -l timezone" actually works but sets it
in smf also) there are still some rough edges.
But just like html, xml is not a proper way to configure systems; it's a
language were was all a customer really wants is a "var=value"; see, e.g.,
/etc/zones/zone.xml vs the way it is configured using zonecfg.
Now, unfortunately, I can't tell exactly what we're changing or when
we are delivering new features. But I will say that clearly heard this
complaint.
>I still live by Solaris (s11) on my laptop and Desktop, sometimes
>painful, but only when I have a config type issue.
>I suspect drivers are going to be harder to come by.
*BSD, etc, and certainly the software for bits found on servers
will be developed and kept up-to-date.
Casper
YTC#1
Mar 27, 2012, 9:17:32 AM3/27/12
to
On 03/27/12 12:41 PM, Casper H.S. Dik wrote:
> YTC#1<b...@ytc1-spambin.co.uk> writes:
>
>> On 03/26/12 09:26 PM, John D Groenveld wrote:
>>> In article<jkq6tg$dil$1...@dont-email.me>, YTC#1<b...@ytc1-spambin.co.uk> wrote:
>>>> And lets not have servers install with NWAM/DHCP by default as well.
>>>
>>> Banged my head against the DHCP (NWAM) selection in the text
>>> installer.
>>> NWAM breaks ipadm(1M):
>>> <URL:http://docs.oracle.com/cd/E23824_01/html/E23811/glmdq.html#glpbq>
>>>
>>> Shame on me for the reading the fucking manual but
>>> Oracle doesn't sell mobile workstation systems and
>>> doesn't seem much interested in supporting a vibrant
>>> third-party Solaris developer ecosystem, so NWAM seems
>>> superfluous.
>
>> I don't think I have ever seen so much swearing in this NG. S11 seems to
>> have upset a lot of people.
>
>> From the workstation/laptop perspective I'd like to hear from Casper,
>> as he has always been an advocate of using Solaris on a laptop.
>
> I'm still using solaris on my laptop.
I never doubted you would be :-)
> YTC#1<b...@ytc1-spambin.co.uk> writes:
>
>> On 03/26/12 09:26 PM, John D Groenveld wrote:
>>> In article<jkq6tg$dil$1...@dont-email.me>, YTC#1<b...@ytc1-spambin.co.uk> wrote:
>>>> And lets not have servers install with NWAM/DHCP by default as well.
>>>
>>> Banged my head against the DHCP (NWAM) selection in the text
>>> installer.
>>> NWAM breaks ipadm(1M):
>>> <URL:http://docs.oracle.com/cd/E23824_01/html/E23811/glmdq.html#glpbq>
>>>
>>> Shame on me for the reading the fucking manual but
>>> Oracle doesn't sell mobile workstation systems and
>>> doesn't seem much interested in supporting a vibrant
>>> third-party Solaris developer ecosystem, so NWAM seems
>>> superfluous.
>
>> I don't think I have ever seen so much swearing in this NG. S11 seems to
>> have upset a lot of people.
>
>> From the workstation/laptop perspective I'd like to hear from Casper,
>> as he has always been an advocate of using Solaris on a laptop.
>
> I'm still using solaris on my laptop.
>
> In Solaris 10 it was even more difficult, especially when you have ethernet
> and wifi. Before that, I needed to had my own "find my location" software
> which barely worked.
I'm thinking of amending it for S11 use on the laptop, and turning NWAM off.
>
> Unfortunately, at the end of the Solaris 11 release some changes were
> made which weren't perhaps complete. While I was able to get the changes to
> setting the zone "fixed" ("zic -l timezone" actually works but sets it
> in smf also) there are still some rough edges.
>
> But just like html, xml is not a proper way to configure systems; it's a
> language were was all a customer really wants is a "var=value"; see, e.g.,
> /etc/zones/zone.xml vs the way it is configured using zonecfg.
>
> Now, unfortunately, I can't tell exactly what we're changing or when
> we are delivering new features. But I will say that clearly heard this
> complaint.
>
>> I still live by Solaris (s11) on my laptop and Desktop, sometimes
>> painful, but only when I have a config type issue.
>
>> I suspect drivers are going to be harder to come by.
>
> Note that for Xorg we're just following what is shipped in Linux
> *BSD, etc, and certainly the software for bits found on servers
> will be developed and kept up-to-date.
OpenSol 134. The graphics then went a bit backwards (can't use an
external screen etc)
Casper H.S. Dik
Mar 27, 2012, 9:32:44 AM3/27/12
to
YTC#1 <b...@ytc1-spambin.co.uk> writes:
>I never could convince you to use the PS UK inetmenu script could I :-).
>I'm thinking of amending it for S11 use on the laptop, and turning NWAM off.
Yeah; I don't remember why I didn't use it but perhaps it
>I never could convince you to use the PS UK inetmenu script could I :-).
>I'm thinking of amending it for S11 use on the laptop, and turning NWAM off.
wasn't magic enough..
>> Note that for Xorg we're just following what is shipped in Linux
>> *BSD, etc, and certainly the software for bits found on servers
>> will be developed and kept up-to-date.
>What about the I915 drivers, I use a Tosh M10 which worked well until
>OpenSol 134. The graphics then went a bit backwards (can't use an
>external screen etc)
Of course, I'm on something using a precursor of Solaris11 update 1
but that shouldn't be surprise.
Casper
John D Groenveld
Mar 27, 2012, 10:42:04 AM3/27/12
to
In article <jks2s8$408$1...@dont-email.me>, YTC#1 <b...@ytc1-spambin.co.uk> wrote:
>I don't think I have ever seen so much swearing in this NG. S11 seems to
>have upset a lot of people.
I don't know what to make of the other person but S11 hasn't
>I don't think I have ever seen so much swearing in this NG. S11 seems to
>have upset a lot of people.
especially upset me.
In total I like it and I intend to continue to use it where
price/performance doesn't require an Illumos based distribution.
The abstraction of the network configuration with dladm(1M),
ipadm(1M), netadm(1M) makes sense if you intend to migrate zones
to different systems, but the NWAM default was a bad idea for
the text installer.
Does Oracle professional services enable it when they set up
the low-volume, high-margin Exa-* appliances for customers?
John
groe...@acm.org
YTC#1
Mar 27, 2012, 12:43:53 PM3/27/12
to
YTC#1
Mar 27, 2012, 12:46:22 PM3/27/12
to
On 03/27/12 02:32 PM, Casper H.S. Dik wrote:
> YTC#1<b...@ytc1-spambin.co.uk> writes:
>
>> I never could convince you to use the PS UK inetmenu script could I :-).
>> I'm thinking of amending it for S11 use on the laptop, and turning NWAM off.
>
> Yeah; I don't remember why I didn't use it but perhaps it
> wasn't magic enough..
It matured, I was using it right up to OpenSol 151
> YTC#1<b...@ytc1-spambin.co.uk> writes:
>
>> I never could convince you to use the PS UK inetmenu script could I :-).
>> I'm thinking of amending it for S11 use on the laptop, and turning NWAM off.
>
> Yeah; I don't remember why I didn't use it but perhaps it
> wasn't magic enough..
>
>>> Note that for Xorg we're just following what is shipped in Linux
>>> *BSD, etc, and certainly the software for bits found on servers
>>> will be developed and kept up-to-date.
>
>> What about the I915 drivers, I use a Tosh M10 which worked well until
>> OpenSol 134. The graphics then went a bit backwards (can't use an
>> external screen etc)
>
> Don't know the details there; what are you corrently using?
(with an SRU)
>
> Of course, I'm on something using a precursor of Solaris11 update 1
> but that shouldn't be surprise.
>
> Casper
Ian Collins
Mar 27, 2012, 2:53:55 PM3/27/12
to
On 03/27/12 10:59 PM, YTC#1 wrote:
> On 03/27/12 05:17 AM, Ian Collins wrote:
>> On 03/27/12 12:41 AM, YTC#1 wrote:
>>> On 03/23/12 03:36 AM, Ian Collins wrote:
>>>> On 03/23/12 10:44 AM, Philip Brown wrote:
>>>>> On Thursday, March 22, 2012 11:13:24 AM UTC-7, Ian Collins wrote:
>>>>>> On 03/23/12 05:25 AM, Philip Brown wrote:
>>>>>>
>>> <snip>
>>>
>>>>> As I have said: not meeting customer needs here.
>>>>
>>>> Maybe (paying) customers simply haven't asked?
>>>>
>>>
>>> They have
>>
>> Care to be more specific?
>>
>
> Like a lot of people I am ex Sun (PS), when S10 appeared and the move
> towards SMF using XML was apparent a lot of people moaned and asked for
> it to be made easier.
>
> We usually just pointed them to the sales/support people. From a PS
> perspective SMF was a (small) money earner.
>
> A PITA though, and I can never remember what bits go where without
> looking at an already working one.
I think the likes of SMF show a programmer's mind set at work rather
> On 03/27/12 05:17 AM, Ian Collins wrote:
>> On 03/27/12 12:41 AM, YTC#1 wrote:
>>> On 03/23/12 03:36 AM, Ian Collins wrote:
>>>> On 03/23/12 10:44 AM, Philip Brown wrote:
>>>>> On Thursday, March 22, 2012 11:13:24 AM UTC-7, Ian Collins wrote:
>>>>>> On 03/23/12 05:25 AM, Philip Brown wrote:
>>>>>>
>>> <snip>
>>>
>>>>> As I have said: not meeting customer needs here.
>>>>
>>>> Maybe (paying) customers simply haven't asked?
>>>>
>>>
>>> They have
>>
>> Care to be more specific?
>>
>
> Like a lot of people I am ex Sun (PS), when S10 appeared and the move
> towards SMF using XML was apparent a lot of people moaned and asked for
> it to be made easier.
>
> We usually just pointed them to the sales/support people. From a PS
> perspective SMF was a (small) money earner.
>
> A PITA though, and I can never remember what bits go where without
> looking at an already working one.
than a system administrator's.
As someone who wears both hats, I'm used to reusing working code, so I
don't find setting up services too much of a pain. But I can appreciate
(more so after this thread!) how others might.
--
Ian Collins
Ian Collins
Mar 27, 2012, 3:11:52 PM3/27/12
to
On 03/27/12 10:56 PM, YTC#1 wrote:
> On 03/26/12 09:26 PM, John D Groenveld wrote:
>> In article<jkq6tg$dil$1...@dont-email.me>, YTC#1<b...@ytc1-spambin.co.uk> wrote:
>>> And lets not have servers install with NWAM/DHCP by default as well.
>>
>> Banged my head against the DHCP (NWAM) selection in the text
>> installer.
>> NWAM breaks ipadm(1M):
>> <URL:http://docs.oracle.com/cd/E23824_01/html/E23811/glmdq.html#glpbq>
>>
>> Shame on me for the reading the fucking manual but
>> Oracle doesn't sell mobile workstation systems and
>> doesn't seem much interested in supporting a vibrant
>> third-party Solaris developer ecosystem, so NWAM seems
>> superfluous.
>
> I don't think I have ever seen so much swearing in this NG. S11 seems to
> have upset a lot of people.
>
> From the workstation/laptop perspective I'd like to hear from Casper,
> as he has always been an advocate of using Solaris on a laptop.
>
> I still live by Solaris (s11) on my laptop and Desktop, sometimes
> painful, but only when I have a config type issue.
Solaris 11 is still an excellent desktop OS, but the absence of third
> On 03/26/12 09:26 PM, John D Groenveld wrote:
>> In article<jkq6tg$dil$1...@dont-email.me>, YTC#1<b...@ytc1-spambin.co.uk> wrote:
>>> And lets not have servers install with NWAM/DHCP by default as well.
>>
>> Banged my head against the DHCP (NWAM) selection in the text
>> installer.
>> NWAM breaks ipadm(1M):
>> <URL:http://docs.oracle.com/cd/E23824_01/html/E23811/glmdq.html#glpbq>
>>
>> Shame on me for the reading the fucking manual but
>> Oracle doesn't sell mobile workstation systems and
>> doesn't seem much interested in supporting a vibrant
>> third-party Solaris developer ecosystem, so NWAM seems
>> superfluous.
>
> I don't think I have ever seen so much swearing in this NG. S11 seems to
> have upset a lot of people.
>
> From the workstation/laptop perspective I'd like to hear from Casper,
> as he has always been an advocate of using Solaris on a laptop.
>
> I still live by Solaris (s11) on my laptop and Desktop, sometimes
> painful, but only when I have a config type issue.
party application repositories (they area Linux has a real edge) may be
the death of of it.
On modern laptops, I just run Solaris full screen in VirtualBox! The
performance on my 2011 Mac is better than it would have been native on a
2010 model. I'm sure the same applies for most other brands as well.
--
Ian Collins
Doug McIntyre
Mar 27, 2012, 3:35:34 PM3/27/12
to
Ian Collins <ian-...@hotmail.com> writes:
>I think the likes of SMF show a programmer's mind set at work rather
>than a system administrator's.
>As someone who wears both hats, I'm used to reusing working code, so I
>don't find setting up services too much of a pain. But I can appreciate
>(more so after this thread!) how others might.
If you take it one step more though, it is a given that a sysadmin
>I think the likes of SMF show a programmer's mind set at work rather
>than a system administrator's.
>As someone who wears both hats, I'm used to reusing working code, so I
>don't find setting up services too much of a pain. But I can appreciate
>(more so after this thread!) how others might.
would know sh shell scripting so they can write init scripts. Back
when SunOS rolled from BSD where a lot of things just launched out of
a one line edit of /etc/rc.local; into Solaris with SVR4 init scripts,
there was a gnashing of teeth and wailing about how much overhead and
stress that caused as well at the time..
Doug McIntyre
Mar 27, 2012, 3:42:51 PM3/27/12
to
Ian Collins <ian-...@hotmail.com> writes:
>Solaris 11 is still an excellent desktop OS, but the absence of third
>party application repositories (they area Linux has a real edge) may be
>the death of of it.
I find that Solaris already has much of what I already want to do
>Solaris 11 is still an excellent desktop OS, but the absence of third
>party application repositories (they area Linux has a real edge) may be
>the death of of it.
packaged up, and the linux repos usually have the kitchen sink needing
to be brought in. Big packages with every database driver loaded,
a desktop GUI option as well as text (ie. mtr) bringing in all of
Gnome, etc.
That many times I build the linux package by hand as well to get rid
of the cruft. The main times I find the linux repo useful is a huge
project that would take a day+ anyway to get all the dependencies
needed to compile it. (ie. xbmc).
But OpenCSW is in process of doing IPS packages for Solaris 11, and
there are things like the OpenIndiana package repo as well, although
they don't seem to do up any package that I actually need in there..
Or in my own repo so I can bring in the packages the way I want them
setup and compiled (although I don't have the structure to keep things
up to date as needed to actually publish that to the general public).
ChrisQ
Mar 27, 2012, 5:35:28 PM3/27/12
to
language (xml) for simple admin task like adding a new service ?.
While it may easy to edit an existing script, it does fall into the admin
hacks category, when the process should be more formalised.
I've bookmarked the manifold link - looks like the sort of utility that Sun
should have provided in the first place, especially if init.d is now
considered
"incorrect" :-)...
Regards,
Chris
Casper H.S. Dik
Mar 28, 2012, 3:26:41 AM3/28/12
to
system wouldn't boot. With init.d that wouldn't happen.
SMF has a lot of advantages; apart from the ability to
enable/disable services which init.d couldn't do (hence checkconf
and other such mechanisms invented for init.d) dependencies can be
set. The newer code in Solaris 11 allows admins to list the changes
made to smf; "svccfg listcust" and for "sys-unconfig" we have
"svccfg delcust".
Casper
YTC#1
Mar 28, 2012, 4:21:12 AM3/28/12
to
:-)
I like VB, but as a tool under my OS. Natively running S11 is far better
(IMO).
> performance on my 2011 Mac is better than it would have been native on a
> 2010 model. I'm sure the same applies for most other brands as well.
>
--
YTC#1
Mar 28, 2012, 4:22:34 AM3/28/12
to
YTC#1
Mar 28, 2012, 4:26:56 AM3/28/12
to
On 03/28/12 08:26 AM, Casper H.S. Dik wrote:
> Doug McIntyre<mer...@geeks.org> writes:
>
>> Ian Collins<ian-...@hotmail.com> writes:
>>> I think the likes of SMF show a programmer's mind set at work rather
>>> than a system administrator's.
>
>>> As someone who wears both hats, I'm used to reusing working code, so I
>>> don't find setting up services too much of a pain. But I can appreciate
>>> (more so after this thread!) how others might.
>
>> If you take it one step more though, it is a given that a sysadmin
>> would know sh shell scripting so they can write init scripts. Back
>> when SunOS rolled from BSD where a lot of things just launched out of
>> a one line edit of /etc/rc.local; into Solaris with SVR4 init scripts,
>> there was a gnashing of teeth and wailing about how much overhead and
>> stress that caused as well at the time..
>
> I remember the downside of rc.local; one syntax error and the
> system wouldn't boot. With init.d that wouldn't happen.
>
> SMF has a lot of advantages; apart from the ability to
> enable/disable services which init.d couldn't do (hence checkconf
I don't think anyone doubts that is/was a great improvement.
> Doug McIntyre<mer...@geeks.org> writes:
>
>> Ian Collins<ian-...@hotmail.com> writes:
>>> I think the likes of SMF show a programmer's mind set at work rather
>>> than a system administrator's.
>
>>> As someone who wears both hats, I'm used to reusing working code, so I
>>> don't find setting up services too much of a pain. But I can appreciate
>>> (more so after this thread!) how others might.
>
>> If you take it one step more though, it is a given that a sysadmin
>> would know sh shell scripting so they can write init scripts. Back
>> when SunOS rolled from BSD where a lot of things just launched out of
>> a one line edit of /etc/rc.local; into Solaris with SVR4 init scripts,
>> there was a gnashing of teeth and wailing about how much overhead and
>> stress that caused as well at the time..
>
> I remember the downside of rc.local; one syntax error and the
> system wouldn't boot. With init.d that wouldn't happen.
>
> SMF has a lot of advantages; apart from the ability to
> enable/disable services which init.d couldn't do (hence checkconf
It is the user interaction (XML) that is the problem. From my experience
I'd say HTML is much easier to code :-)
A few years back I was working on a project using zones and there was a
push to get N1 involved. I was asked to use it on the project and be the
PRE for it.
After being shown round the product and all the XML to do what were
simple tasks when scripted, I asked to be taken off the project to get
as far away from XML as possible :-)
> and other such mechanisms invented for init.d) dependencies can be
> set. The newer code in Solaris 11 allows admins to list the changes
> made to smf; "svccfg listcust" and for "sys-unconfig" we have
> "svccfg delcust".
>
> Casper
Ian Collins
Mar 28, 2012, 4:28:20 AM3/28/12
to
On 03/28/12 09:21 PM, YTC#1 wrote:
> On 03/27/12 08:11 PM, Ian Collins wrote:
>>
>> On modern laptops, I just run Solaris full screen in VirtualBox! The
>
> Which means you don't run S11 on your laptop.
> :-)
But if fools most people most of the time!
> On 03/27/12 08:11 PM, Ian Collins wrote:
>>
>> On modern laptops, I just run Solaris full screen in VirtualBox! The
>
> Which means you don't run S11 on your laptop.
> :-)
> I like VB, but as a tool under my OS. Natively running S11 is far better
> (IMO).
could get it running, I'd loose dynamic graphics switching which I use a
lot.
--
Ian Collins
YTC#1
Mar 28, 2012, 5:31:27 AM3/28/12
to
cheerfull AMD/intel based laptops.
Philip Brown
Mar 28, 2012, 1:37:16 PM3/28/12
to
On Wednesday, March 28, 2012 1:26:56 AM UTC-7, YTC#1 wrote:
> On 03/28/12 08:26 AM, Casper H.S. Dik wrote:
>
> > SMF has a lot of advantages; apart from the ability to
> > enable/disable services which init.d couldn't do (hence checkconf
>
> I don't think anyone doubts that is/was a great improvement.
>
> It is the user interaction (XML) that is the problem. From my experience
> I'd say HTML is much easier to code :-)
>
Exactly.
> On 03/28/12 08:26 AM, Casper H.S. Dik wrote:
>
> > SMF has a lot of advantages; apart from the ability to
> > enable/disable services which init.d couldn't do (hence checkconf
>
> I don't think anyone doubts that is/was a great improvement.
>
> It is the user interaction (XML) that is the problem. From my experience
> I'd say HTML is much easier to code :-)
>
It is insane, that Oracle thinks sysadmins now need to be forced to use a tool, for something as "simple" as configuring a network interface... what used to be one line in /etc/hostname.XYZ0 ...
Yet, at the same time, it's just fine for us to have to hand-edit
**100-line** xml files for SMF?
That's nuts.
kartikvashishta108
Mar 30, 2012, 9:47:58 AM3/30/12
to
Apple thrives on. I have followed intel/amd laptop/desktop vs apple
system prices over the years and I can say witth authority that Apple
products are marked more expensive thereby denying them to
minorities(usually poor) and also to those others who can't afford
them. Same applies to gadgets like iPod - I'll go with a $30 mp3
player and iPad - I'd go with a netbook.
Regards,
Kartik Vashishta
kartikvashishta108
Mar 30, 2012, 3:47:09 PM3/30/12
to
http://www.oracle.com/technetwork/server-storage/solaris/solaris-smf-wp-167901.pdf
some interesting links are at the end of the whitepaper.
chuckers
Apr 1, 2012, 8:23:48 PM4/1/12
to
On Friday, March 30, 2012 10:47:58 PM UTC+9, kartikvashishta108 wrote:
[edit]
Racism is what is making Apple kit "expensive"?
You are an authority on this? Really?
Expensive products are denied to minorities?
[edit]
> It is actually racism and depriving the poor people of technology that
> Apple thrives on. I have followed intel/amd laptop/desktop vs apple
> system prices over the years and I can say witth authority that Apple
> products are marked more expensive thereby denying them to
> minorities(usually poor) and also to those others who can't afford
> them. Same applies to gadgets like iPod - I'll go with a $30 mp3
> player and iPad - I'd go with a netbook.
>
Wow, that is a whole lot of crap packed into 3 sentences.
> Apple thrives on. I have followed intel/amd laptop/desktop vs apple
> system prices over the years and I can say witth authority that Apple
> products are marked more expensive thereby denying them to
> minorities(usually poor) and also to those others who can't afford
> them. Same applies to gadgets like iPod - I'll go with a $30 mp3
> player and iPad - I'd go with a netbook.
>
Racism is what is making Apple kit "expensive"?
You are an authority on this? Really?
Expensive products are denied to minorities?
kartikvashishta108
Apr 2, 2012, 8:02:43 AM4/2/12
to
chuckers
Apr 2, 2012, 7:25:43 PM4/2/12
to
http://go.bloomberg.com/tech-blog/2012-03-20-now-can-we-start-talking-about-the-real-foxconn/
You still haven't supported your thesis: "racism and depriving the poor people of technology that Apple thrives on"
kartikvashishta108
Apr 2, 2012, 9:28:56 PM4/2/12
to
On Apr 2, 6:25 pm, chuckers <chucker...@gmail.com> wrote:
> On Monday, April 2, 2012 9:02:43 PM UTC+9, kartikvashishta108 wrote:
> > On Apr 1, 7:23 pm, chuckers <chucker...@gmail.com> wrote:
> > > On Friday, March 30, 2012 10:47:58 PM UTC+9, kartikvashishta108 wrote:
>
> > > [edit]
>
> > > > It is actually racism and depriving the poor people of technology that
> > > > Apple thrives on. I have followed intel/amd laptop/desktop vs apple
> > > > system prices over the years and I can say witth authority that Apple
> > > > products are marked more expensive thereby denying them to
> > > > minorities(usually poor) and also to those others who can't afford
> > > > them. Same applies to gadgets like iPod - I'll go with a $30 mp3
> > > > player and iPad - I'd go with a netbook.
>
> > > Wow, that is a whole lot of crap packed into 3 sentences.
>
> > > Racism is what is making Apple kit "expensive"?
> > > You are an authority on this? Really?
> > > Expensive products are denied to minorities?
>
> >http://www.guardian.co.uk/technology/2011/apr/30/apple-chinese-factor...
> On Monday, April 2, 2012 9:02:43 PM UTC+9, kartikvashishta108 wrote:
> > On Apr 1, 7:23 pm, chuckers <chucker...@gmail.com> wrote:
> > > On Friday, March 30, 2012 10:47:58 PM UTC+9, kartikvashishta108 wrote:
>
> > > [edit]
>
> > > > It is actually racism and depriving the poor people of technology that
> > > > Apple thrives on. I have followed intel/amd laptop/desktop vs apple
> > > > system prices over the years and I can say witth authority that Apple
> > > > products are marked more expensive thereby denying them to
> > > > minorities(usually poor) and also to those others who can't afford
> > > > them. Same applies to gadgets like iPod - I'll go with a $30 mp3
> > > > player and iPad - I'd go with a netbook.
>
> > > Wow, that is a whole lot of crap packed into 3 sentences.
>
> > > Racism is what is making Apple kit "expensive"?
> > > You are an authority on this? Really?
> > > Expensive products are denied to minorities?
>
>
> Ooohh...I can put up a link too:
>
> http://go.bloomberg.com/tech-blog/2012-03-20-now-can-we-start-talking...
> Ooohh...I can put up a link too:
>
>
> You still haven't supported your thesis: "racism and depriving the poor people of technology that Apple thrives on"
Depriving minorities(usually poor) - not that I'm saying there's no
> You still haven't supported your thesis: "racism and depriving the poor people of technology that Apple thrives on"
one in the "majority"who's not poor is an instance of racism. Apple
had/has good technology, but it is consistently priced out of the
reach of certain segments of society.
Opensolaris, solaris itself, BSD(great example of being consistently
free and available to ALL - on cheap platforms), Linux etc are
examples of non-racist technology.
There's "racist" consequences of pricing technology.
Ian Collins
Apr 2, 2012, 11:30:05 PM4/2/12
to
On 03/31/12 02:47 AM, kartikvashishta108 wrote:
> On Mar 28, 4:31 am, YTC#1<b...@ytc1-spambin.co.uk> wrote:
>> On 03/28/12 09:28 AM, Ian Collins wrote:
>>> On 03/28/12 09:21 PM, YTC#1 wrote:
>>>> On 03/27/12 08:11 PM, Ian Collins wrote:
>>
>>>>> On modern laptops, I just run Solaris full screen in VirtualBox! The
>>
>>>> Which means you don't run S11 on your laptop.
>>>> :-)
>>
>>> But if fools most people most of the time!
>>
>>>> I like VB, but as a tool under my OS. Natively running S11 is far better
>>>> (IMO).
>>
>>> But it's a bit of a pain (if even possible) on current Macs, Even if I
>>> could get it running, I'd loose dynamic graphics switching which I use a
>>> lot.
>>
>> Not being a MACisti I don't get that pain, I stick with cheap(ish) and
>> cheerfull AMD/intel based laptops.
>>
The least you could have done was snip the sig block before launching
> On Mar 28, 4:31 am, YTC#1<b...@ytc1-spambin.co.uk> wrote:
>> On 03/28/12 09:28 AM, Ian Collins wrote:
>>> On 03/28/12 09:21 PM, YTC#1 wrote:
>>>> On 03/27/12 08:11 PM, Ian Collins wrote:
>>
>>>>> On modern laptops, I just run Solaris full screen in VirtualBox! The
>>
>>>> Which means you don't run S11 on your laptop.
>>>> :-)
>>
>>> But if fools most people most of the time!
>>
>>>> I like VB, but as a tool under my OS. Natively running S11 is far better
>>>> (IMO).
>>
>>> But it's a bit of a pain (if even possible) on current Macs, Even if I
>>> could get it running, I'd loose dynamic graphics switching which I use a
>>> lot.
>>
>> Not being a MACisti I don't get that pain, I stick with cheap(ish) and
>> cheerfull AMD/intel based laptops.
>>
into your off-topic rant.
> It is actually racism and depriving the poor people of technology that
> Apple thrives on. I have followed intel/amd laptop/desktop vs apple
> system prices over the years and I can say witth authority that Apple
> products are marked more expensive thereby denying them to
> minorities(usually poor) and also to those others who can't afford
> them. Same applies to gadgets like iPod - I'll go with a $30 mp3
> player and iPad - I'd go with a netbook.
So every luxury goods maker is racist? Wow. I guess that makes oracle
racist as well?
--
Ian Collins
chuckers
Apr 3, 2012, 2:12:47 AM4/3/12
to
Minorities are NOT being deprived of Apple products because of racism. Anyone is able to make use of the products that can afford them. Being poor is an equal opportunity problem. You are equating minorities with those poor. That is demonstrably incorrect.
Are you saying that the makers of sports cars are racists because the poor can't afford them?
What about grocery stores that refuse to sell their goods to those that can't afford them? Are they racist too?
You overuse and INCORRECT use of the word "racism" only serves to devalue the words and doing that only encourages real racism to flourish. I suggest you go find a dictionary and learn what the word actually means.
kartikvashishta108
Apr 3, 2012, 9:31:59 AM4/3/12
to
"certain textbooks" are priced out of the reach of those who are not
able to afford them. Hence depriving those people of the ability to
learn from those certain "textbooks". It is a rather subtle issue.
YTC#1
Apr 3, 2012, 2:46:01 PM4/3/12
to
On 28/03/2012 18:37, Philip Brown wrote:
> On Wednesday, March 28, 2012 1:26:56 AM UTC-7, YTC#1 wrote:
>> On 03/28/12 08:26 AM, Casper H.S. Dik wrote:
>>
>>> SMF has a lot of advantages; apart from the ability to
>>> enable/disable services which init.d couldn't do (hence checkconf
>>
>> I don't think anyone doubts that is/was a great improvement.
>>
>> It is the user interaction (XML) that is the problem. From my experience
>> I'd say HTML is much easier to code :-)
>>
>
>
> Exactly.
>
> It is insane, that Oracle thinks sysadmins now need to be forced to use a tool, for something as "simple" as configuring a network interface... what used to be one line in /etc/hostname.XYZ0 ...
>
Worse than that, it was Sun when the slide started :-(
> On Wednesday, March 28, 2012 1:26:56 AM UTC-7, YTC#1 wrote:
>> On 03/28/12 08:26 AM, Casper H.S. Dik wrote:
>>
>>> SMF has a lot of advantages; apart from the ability to
>>> enable/disable services which init.d couldn't do (hence checkconf
>>
>> I don't think anyone doubts that is/was a great improvement.
>>
>> It is the user interaction (XML) that is the problem. From my experience
>> I'd say HTML is much easier to code :-)
>>
>
>
> Exactly.
>
> It is insane, that Oracle thinks sysadmins now need to be forced to use a tool, for something as "simple" as configuring a network interface... what used to be one line in /etc/hostname.XYZ0 ...
>
YTC#1
Apr 3, 2012, 2:49:23 PM4/3/12
to
> "certain textbooks" are priced out of the reach of those who are not
> able to afford them. Hence depriving those people of the ability to
> learn from those certain "textbooks". It is a rather subtle issue.
Ian Collins
Apr 3, 2012, 3:15:04 PM4/3/12
to
On 04/ 4/12 06:46 AM, YTC#1 wrote:
> On 28/03/2012 18:37, Philip Brown wrote:
>> On Wednesday, March 28, 2012 1:26:56 AM UTC-7, YTC#1 wrote:
>>> On 03/28/12 08:26 AM, Casper H.S. Dik wrote:
>>>
>>>> SMF has a lot of advantages; apart from the ability to
>>>> enable/disable services which init.d couldn't do (hence checkconf
>>>
>>> I don't think anyone doubts that is/was a great improvement.
>>>
>>> It is the user interaction (XML) that is the problem. From my experience
>>> I'd say HTML is much easier to code :-)
>>>
>>
>>
>> Exactly.
>>
>> It is insane, that Oracle thinks sysadmins now need to be forced to use a tool, for something as "simple" as configuring a network interface... what used to be one line in /etc/hostname.XYZ0 ...
>>
>
> Worse than that, it was Sun when the slide started :-(
Recently I've been working with some Solaris 11 systems with mix of
> On 28/03/2012 18:37, Philip Brown wrote:
>> On Wednesday, March 28, 2012 1:26:56 AM UTC-7, YTC#1 wrote:
>>> On 03/28/12 08:26 AM, Casper H.S. Dik wrote:
>>>
>>>> SMF has a lot of advantages; apart from the ability to
>>>> enable/disable services which init.d couldn't do (hence checkconf
>>>
>>> I don't think anyone doubts that is/was a great improvement.
>>>
>>> It is the user interaction (XML) that is the problem. From my experience
>>> I'd say HTML is much easier to code :-)
>>>
>>
>>
>> Exactly.
>>
>> It is insane, that Oracle thinks sysadmins now need to be forced to use a tool, for something as "simple" as configuring a network interface... what used to be one line in /etc/hostname.XYZ0 ...
>>
>
> Worse than that, it was Sun when the slide started :-(
network ports and configurations and once I got used to the new
commands, they do make life easier. Compare setting up IPMP on Solaris
10 with Solaris 11 for example.
--
Ian Collins
chuckers
Apr 3, 2012, 7:09:25 PM4/3/12
to
Cydrome Leader
Apr 3, 2012, 10:57:50 PM4/3/12
to
Jim Leonard
Apr 5, 2012, 4:33:09 PM4/5/12
to
On Apr 3, 9:57 pm, Cydrome Leader <prese...@MUNGEpanix.com> wrote:
> IPMP is complete garbage is the problem with IPMP.
What are your preferred alternatives?
> IPMP is complete garbage is the problem with IPMP.
Jim Leonard
Apr 5, 2012, 4:36:52 PM4/5/12
to
On Apr 3, 8:31 am, kartikvashishta108 <kartik.u...@gmail.com> wrote:
> You guys are comparing apples to oranges. A close analogy is that
> "certain textbooks" are priced out of the reach of those who are not
> able to afford them. Hence depriving those people of the ability to
> learn from those certain "textbooks". It is a rather subtle issue.
There is nothing subtle about the concept of racism. Racism is
> You guys are comparing apples to oranges. A close analogy is that
> "certain textbooks" are priced out of the reach of those who are not
> able to afford them. Hence depriving those people of the ability to
> learn from those certain "textbooks". It is a rather subtle issue.
discrimination based on race -- that's it. I'm not sure what your
definition is, but it's incorrect.
You could argue that expensive products are discrimination based on
wealth, sure. But not race.
Cydrome Leader
Apr 6, 2012, 1:34:51 AM4/6/12
to
Jim Leonard <moby...@gmail.com> wrote:
It's a standard and it's non-retarded as it wasn't concocted by sun.
dladm create-aggr
will get you started with it.
We ran IPMP for a while. that just resulted in stuff being broken all the
time, and lots of reboots to unwedge botched up links. Even on the older
sparc machines (that I believe don't support LACP) IPMP always caused more
problems than it solved. It always errored out more often than ethernet
cables failed, NICs broke or ports on a cisco switch blew out.
After using LACP on the same machines, we forgot about all ever losing
links to machines. your switches need to support it, but any real switch
does, so that's not a problem.
IPMP is as jenky as those old putty colored metal boxes for switching
parallel ports between computers and printers.
> On Apr 3, 9:57?pm, Cydrome Leader <prese...@MUNGEpanix.com> wrote:
>> IPMP is complete garbage is the problem with IPMP.
>
> What are your preferred alternatives?
LACP.
>> IPMP is complete garbage is the problem with IPMP.
>
> What are your preferred alternatives?
It's a standard and it's non-retarded as it wasn't concocted by sun.
dladm create-aggr
will get you started with it.
We ran IPMP for a while. that just resulted in stuff being broken all the
time, and lots of reboots to unwedge botched up links. Even on the older
sparc machines (that I believe don't support LACP) IPMP always caused more
problems than it solved. It always errored out more often than ethernet
cables failed, NICs broke or ports on a cisco switch blew out.
After using LACP on the same machines, we forgot about all ever losing
links to machines. your switches need to support it, but any real switch
does, so that's not a problem.
IPMP is as jenky as those old putty colored metal boxes for switching
parallel ports between computers and printers.
Ian Collins
Apr 6, 2012, 1:56:57 AM4/6/12
to
On 04/ 6/12 05:34 PM, Cydrome Leader wrote:
> Jim Leonard<moby...@gmail.com> wrote:
>> On Apr 3, 9:57?pm, Cydrome Leader<prese...@MUNGEpanix.com> wrote:
>>> IPMP is complete garbage is the problem with IPMP.
>>
>> What are your preferred alternatives?
>
> LACP.
>
> It's a standard and it's non-retarded as it wasn't concocted by sun.
>
> dladm create-aggr
>
> will get you started with it.
I agree.
> Jim Leonard<moby...@gmail.com> wrote:
>> On Apr 3, 9:57?pm, Cydrome Leader<prese...@MUNGEpanix.com> wrote:
>>> IPMP is complete garbage is the problem with IPMP.
>>
>> What are your preferred alternatives?
>
> LACP.
>
> It's a standard and it's non-retarded as it wasn't concocted by sun.
>
> dladm create-aggr
>
> will get you started with it.
It is also another example of the simplification of configuration with
the Solaris 11 commands!
--
Ian Collins
Richard B. Gilbert
Apr 6, 2012, 7:35:20 PM4/6/12
to
cindy
Apr 11, 2012, 1:52:41 PM4/11/12
to
On Mar 21, 1:26 pm, cindy <cindy.swearin...@oracle.com> wrote:
> On Mar 21, 12:16 pm, Philip Brown <p...@bolthole.com> wrote:
>
> > I'm not a big redhat fan, but I bring up something they do as a point of reference.
>
> > Turning (random process) into a boot-time service, is trivial.
>
> > 1. Make a script for it in /etc/init.d that supports "start" and "stop" options.
> > 2, run "chkconfig (newname) on"
>
> > DONE.
>
> > On solaris, it is painful. You still have part 1, make an init script... but then the docs tell you to wade through a mountain of xml. Yuck!
>
> > Is there no common "make an SMF service" helper out there yet? My searches did not yield fruit.
> > It would be really surprising if there still isnt one yet, after 7 years of SMF
>
> Hi Phil,
>
> Check out pages 49-51 of this doc:
>
> http://timsfoster.files.wordpress.com/2011/10/book.pdf
A follow up is that I should have looked a little harder.
This issue is already filed and is being actively worked:
7125253 Need a tool to make it easier to generate SMF manifests
Thanks,
Cindy
> On Mar 21, 12:16 pm, Philip Brown <p...@bolthole.com> wrote:
>
> > I'm not a big redhat fan, but I bring up something they do as a point of reference.
>
> > Turning (random process) into a boot-time service, is trivial.
>
> > 1. Make a script for it in /etc/init.d that supports "start" and "stop" options.
> > 2, run "chkconfig (newname) on"
>
> > DONE.
>
> > On solaris, it is painful. You still have part 1, make an init script... but then the docs tell you to wade through a mountain of xml. Yuck!
>
> > Is there no common "make an SMF service" helper out there yet? My searches did not yield fruit.
> > It would be really surprising if there still isnt one yet, after 7 years of SMF
>
> Hi Phil,
>
> Check out pages 49-51 of this doc:
>
> http://timsfoster.files.wordpress.com/2011/10/book.pdf
A follow up is that I should have looked a little harder.
This issue is already filed and is being actively worked:
7125253 Need a tool to make it easier to generate SMF manifests
Thanks,
Cindy
Philip Brown
Apr 13, 2012, 8:23:44 PM4/13/12
to
On Wednesday, April 11, 2012 10:52:41 AM UTC-7, cindy wrote:
>
>
> A follow up is that I should have looked a little harder.
> This issue is already filed and is being actively worked:
>
> 7125253 Need a tool to make it easier to generate SMF manifests
>
Thank you Cindy.
>
>
> A follow up is that I should have looked a little harder.
> This issue is already filed and is being actively worked:
>
> 7125253 Need a tool to make it easier to generate SMF manifests
>
It's nice to note that in the bugid, a fix was referenced as being coded, and in the works for release.
Fixed Version s11u1_14
It would be nice to know the command name that implements
"PSARC 2012/087 SMF Manifest Generation Tool."
so we can go hunt for it in the updates
cindy swearingen
Apr 16, 2012, 11:04:38 AM4/16/12
to
Keep in mind too that Oracle provides Solaris beta programs so that
you
can access and evaluate new features.
Thanks,
Cindy
David Combs
Apr 21, 2012, 11:39:52 PM4/21/12
to
In article <9t29ds...@mid.individual.net>,
Ian Collins <ian-...@hotmail.com> wrote:
Although I must admit I have a replication
>service I often add to new machines or zones and I have written a simple
>tool to generate host specific manifests.
>
>--
>Ian Collins
This tool you've written -- maybe show it here, with hints
of how to create similar ones for other like tasks?
David
Ian Collins <ian-...@hotmail.com> wrote:
Although I must admit I have a replication
>service I often add to new machines or zones and I have written a simple
>tool to generate host specific manifests.
>
>--
>Ian Collins
This tool you've written -- maybe show it here, with hints
of how to create similar ones for other like tasks?
David
David Combs
May 5, 2012, 4:54:41 PM5/5/12
to
In article <jkq6tg$dil$1...@dont-email.me>, YTC#1 <b...@ytc1-spambin.co.uk> wrote:
>On 03/26/12 04:47 PM, Rich Teer wrote:
>> On Mon, 26 Mar 2012, YTC#1 wrote:
>>
>>> /etc/hostname.<nic> is a config file
>>>
>>> That was a nice easy thing to use in SunOS until S11
>>
>> I must admit (and it seems I'm not alone here), but as Solaris
>> moves further and further away from plain old text config files
>> (in the grand old UNIX tradition), us old timers appreciate it
>
>I don't class myself as old, yet :-)
>
>> less and less. What next, a registry a la Windoze?!
>
>That has been an ugly rumour that has circulated for a long time....
>
>>
>> Give me plain text files and vi over n (for suitably large values
>> of n) bespoke tools for editing opaque XML (or other) files!
>>
>
>+1
>
>And lets not have servers install with NWAM/DHCP by default as well.
And I suppose, at least with emacs, there could be
various modes for editing the more complex ones.
Maybe even with some protection built-in for
catching "stupid" errors, like entering unlikely values,
things like that.
OK -- about Solaris seemingly moving away from ASCII
config files:
WHO is it who decides on such things?
Maybe some of the more knowledgeable in this group, say,
could work to become outside advisors to or members
of whatever groups decide this stuff.
Likely the MBA-types will insist on GUI-type interfaces,
a la Microsoft.
BUT, even so, maybe at very low cost they could make available
a parallel text-based interface.
Plus, for even those who prefer the GUI style, once the
config for something has been set up, maybe you could
talk the Oracle (or whatever) people into a facility
that would take that config-info and write it out
in an ASCII file (such that it can be read back in).
The sales point of course being that it makes it
simpler to config multiple machines. And if
the config can vary slightly from one machine to
another, it's certainly simple enough to make
minor edits for each one. Even via sed! (via
saved sed-files, one per different config)
Also, easier training of new unix sysadmins, since
each will ALREADY be proficient in these text-based
tools.
And, if you can get your company behind you, and
it's big enough to be important to Oracle or whoever,
they can threaten to to linix. Pretty sure they
won't think you're bluffing.
Far better to make the effort than to just
sit back and let the marketing types decide
what's going to be inflicted on you!
David
PS: No, I can't do it.
A: I'm just a sole operator.
B: From my question-posts here, you know only
too well how technically IGNORANT I am
on sysadmin stuff.
>On 03/26/12 04:47 PM, Rich Teer wrote:
>> On Mon, 26 Mar 2012, YTC#1 wrote:
>>
>>> /etc/hostname.<nic> is a config file
>>>
>>> That was a nice easy thing to use in SunOS until S11
>>
>> I must admit (and it seems I'm not alone here), but as Solaris
>> moves further and further away from plain old text config files
>> (in the grand old UNIX tradition), us old timers appreciate it
>
>I don't class myself as old, yet :-)
>
>> less and less. What next, a registry a la Windoze?!
>
>That has been an ugly rumour that has circulated for a long time....
>
>>
>> Give me plain text files and vi over n (for suitably large values
>> of n) bespoke tools for editing opaque XML (or other) files!
>>
>
>+1
>
>And lets not have servers install with NWAM/DHCP by default as well.
>
>
>
>--
>Bruce Porter
>
>"The internet is a huge and diverse community but mainly friendly"
>http://blog.maui.co.uk/index.php/ytc/
>There *is* an alternative! http://www.openoffice.org/
Agreed -- ASCII files are nice for config files.
>
>
>--
>Bruce Porter
>
>"The internet is a huge and diverse community but mainly friendly"
>http://blog.maui.co.uk/index.php/ytc/
>There *is* an alternative! http://www.openoffice.org/
And I suppose, at least with emacs, there could be
various modes for editing the more complex ones.
Maybe even with some protection built-in for
catching "stupid" errors, like entering unlikely values,
things like that.
OK -- about Solaris seemingly moving away from ASCII
config files:
WHO is it who decides on such things?
Maybe some of the more knowledgeable in this group, say,
could work to become outside advisors to or members
of whatever groups decide this stuff.
Likely the MBA-types will insist on GUI-type interfaces,
a la Microsoft.
BUT, even so, maybe at very low cost they could make available
a parallel text-based interface.
Plus, for even those who prefer the GUI style, once the
config for something has been set up, maybe you could
talk the Oracle (or whatever) people into a facility
that would take that config-info and write it out
in an ASCII file (such that it can be read back in).
The sales point of course being that it makes it
simpler to config multiple machines. And if
the config can vary slightly from one machine to
another, it's certainly simple enough to make
minor edits for each one. Even via sed! (via
saved sed-files, one per different config)
Also, easier training of new unix sysadmins, since
each will ALREADY be proficient in these text-based
tools.
And, if you can get your company behind you, and
it's big enough to be important to Oracle or whoever,
they can threaten to to linix. Pretty sure they
won't think you're bluffing.
Far better to make the effort than to just
sit back and let the marketing types decide
what's going to be inflicted on you!
David
PS: No, I can't do it.
A: I'm just a sole operator.
B: From my question-posts here, you know only
too well how technically IGNORANT I am
on sysadmin stuff.
Ian Collins
May 5, 2012, 6:27:37 PM5/5/12
to
On 05/ 6/12 08:54 AM, David Combs wrote:
>
> OK -- about Solaris seemingly moving away from ASCII
> config files:
>
> WHO is it who decides on such things?
Oracle, who else?
>
> OK -- about Solaris seemingly moving away from ASCII
> config files:
>
> WHO is it who decides on such things?
> Maybe some of the more knowledgeable in this group, say,
> could work to become outside advisors to or members
> of whatever groups decide this stuff.
>
> Likely the MBA-types will insist on GUI-type interfaces,
> a la Microsoft.
is the path they are following.
> BUT, even so, maybe at very low cost they could make available
> a parallel text-based interface.
from a central application). Text files are a pain to manage from scripts.
--
Ian Collins
Doug McIntyre
May 5, 2012, 11:49:41 PM5/5/12
to
Ian Collins <ian-...@hotmail.com> writes:
>On 05/ 6/12 08:54 AM, David Combs wrote:
>>
>> OK -- about Solaris seemingly moving away from ASCII
>> config files:
>>
>> WHO is it who decides on such things?
>Oracle, who else?
This was long before Oracle was in the picture. And likely it was from
>On 05/ 6/12 08:54 AM, David Combs wrote:
>>
>> OK -- about Solaris seemingly moving away from ASCII
>> config files:
>>
>> WHO is it who decides on such things?
>Oracle, who else?
customer demand, driven up from the support department with real-life
situations, or from an admin deploying 1,000 unit installs for a large
deployment wishing for such a thing.
>> BUT, even so, maybe at very low cost they could make available
>> a parallel text-based interface.
>The intent appears to be an interface that is easy to script (and manage
>from a central application). Text files are a pain to manage from scripts.
correctly configure from scripting. I've posted this before, but for
our day-to-day ops, we've had to manage text files with scripts, with
multi-admins doing commits on. Its amazing what different things
others will do that the script writer doesn't anticipate them doing
& messing it up.
Its not like a programmer environment, where the individual programmer
has to make it compile first before checkin. In an admin setup, it'll
most likely will fail something production due to a misedit.
Having a programatic interface to configure something like like NIC
settings or physical setup like LAG groups is very very nice.
ChrisQ
May 7, 2012, 7:10:46 AM5/7/12
to
On 05/06/12 03:49, Doug McIntyre wrote:
>
> Exactly. Text based config files are a major pain to verify and
> correctly configure from scripting. I've posted this before, but for
> our day-to-day ops, we've had to manage text files with scripts, with
> multi-admins doing commits on. Its amazing what different things
> others will do that the script writer doesn't anticipate them doing
> & messing it up.
>
> Its not like a programmer environment, where the individual programmer
> has to make it compile first before checkin. In an admin setup, it'll
> most likely will fail something production due to a misedit.
>
> Having a programatic interface to configure something like like NIC
> settings or physical setup like LAG groups is very very nice.
>
It probably doesn't matter what the data format is. The real issue is
>
> Exactly. Text based config files are a major pain to verify and
> correctly configure from scripting. I've posted this before, but for
> our day-to-day ops, we've had to manage text files with scripts, with
> multi-admins doing commits on. Its amazing what different things
> others will do that the script writer doesn't anticipate them doing
> & messing it up.
>
> Its not like a programmer environment, where the individual programmer
> has to make it compile first before checkin. In an admin setup, it'll
> most likely will fail something production due to a misedit.
>
> Having a programatic interface to configure something like like NIC
> settings or physical setup like LAG groups is very very nice.
>
really about trackability and the only way to do that is to put all the
data under some sort of revision control, just as one would do for
s/w projects. I guess it's easier to integrate that sort of thing within
a gui based tool, than at the command line, where the process is more
fragmented in terms of the tools used and their interdependencies...
Regards,
Chris
Cydrome Leader
May 7, 2012, 3:08:51 PM5/7/12
to
deal with than text files.
I can whip text files into shape with sed and awk. Good luck with that on
openview or tivoli.
David Combs
May 13, 2012, 10:17:11 PM5/13/12
to
In article <a0lnqp...@mid.individual.net>,
Wow! Looks like I need some re-education!
Updating multiple machines is something I know nothing about, being
as I have only one Sun (sparc) up and running.
Please, what are some of these non-text-file-based facilities that
you and others here use to update hundreds of machines?
And suppose some of them are to have slight modifications to their
setup? Like if its name has a "foo-" prefix, do this, else do that.
Seems like an inherently programming-like problem, like where you
have to tell the facility to do something it might never have
done before, and probably is not already there as a built-in choice.
Needs some kind of script, I'd think.
So, how do these modern (gui-based?) facilities you guys use handle
such requirements?
I bet that with your reply (replies), my brain is going to
get streched a bit!
Thanks!
David
Updating multiple machines is something I know nothing about, being
as I have only one Sun (sparc) up and running.
Please, what are some of these non-text-file-based facilities that
you and others here use to update hundreds of machines?
And suppose some of them are to have slight modifications to their
setup? Like if its name has a "foo-" prefix, do this, else do that.
Seems like an inherently programming-like problem, like where you
have to tell the facility to do something it might never have
done before, and probably is not already there as a built-in choice.
Needs some kind of script, I'd think.
So, how do these modern (gui-based?) facilities you guys use handle
such requirements?
I bet that with your reply (replies), my brain is going to
get streched a bit!
Thanks!
David
David Combs
May 13, 2012, 10:23:59 PM5/13/12
to
In article <4fa5f4d5$0$96705$8046...@newsreader.iphouse.net>,
When I google LAG groups I'm not sure I'm getting the right thing.
Perhaps you could give me some longer names (ie more unique) that
I can google for and hopefully find nice wikipedia articles on?
THANKS!
David
Perhaps you could give me some longer names (ie more unique) that
I can google for and hopefully find nice wikipedia articles on?
THANKS!
David
David Combs
May 13, 2012, 10:37:02 PM5/13/12
to
In article <jkm05v$eqm$1...@reader1.panix.com>,
Cydrome Leader <pres...@MUNGEpanix.com> wrote:
>Philip Brown <ph...@bolthole.com> wrote:
>> On Thursday, March 22, 2012 1:01:50 AM UTC-7, Casper H. S. Dik wrote:
>>> Philip Brown
>>> service XML file and change the name, dependencies, etc.
>>> Import them (svcadm restart manifest-import)
>>
>> that is semantically almost identical to,
>>
>> "It's not that hard to customise this program. Extract the source code, change the #defines in the right places, and type 'make'"
>>
>> That also falls in the realm of "not that difficult".
>> But it's not a very good 'customer interface' either.
>>
>>> (Clearly, this should be made simpler)
>>
>> I think this is the important take-away from this :) It's long past time that sun^H^H^Horacle provided a decent end-user tool.
>>
>> "Edit the xml file" is NOT "a decent end-user tool".
>> Especially one as convoluted as an SMF manifest.
>>
>> Take a look at the user-level questions that
>> http://code.google.com/p/manifold/
>> asks when you run it.
>>
>> Manifests could have an initial section of a few lines that hold those values, at the top. Rather than having to go through the entire file, "parsing" the xml in your head.
>> If xml is sooo simple to parse, why is libxml2.so 2 megabytes! :P
>> It's not nice to force humans to do something, that takes a machine many many lines of code to do.
>>
>> ugh.
>>
>> Personally, I would prefer a tool even simpler than the "manifold" tool. Something as simple as the redhat "chkconfig" tool.
>> That is to say, default to handling the 95% of most sysadmin's needs, with a minimum of unneccessary questions.
>>
>> It's almost always just a simple case of, "hey, I want this demon started at boot time, here's the start/stop script".
>>
>> Oracle is not meeting customer needs well in this area.
>
>I think it's safe to say solaris is dead this point.
>
>good job oracle.
>
>
So, what did they pay all that money for? Just patents?
(Against google, anyway, they seem to have lost their
gamble within the last week or two, getting probably
awarded 1% or less of what they expected. Or did
I read the newspaper wrong?)
If solaris "is dead at this point", then what are they going to do
for an OS? Linux on x86? I thought Suns were supposed to be
so much better architected for running Oracle databases?
Someone please comment on this "solaris is dead at this point"
statement.
THANKS!
David
Cydrome Leader <pres...@MUNGEpanix.com> wrote:
>Philip Brown <ph...@bolthole.com> wrote:
>> On Thursday, March 22, 2012 1:01:50 AM UTC-7, Casper H. S. Dik wrote:
>>> Philip Brown
>>>
>>> >Is there no common "make an SMF service" helper out there yet? My searches did not yield fruit.
>>> >It would be really surprising if there still isnt one yet, after 7 years of SMF
>>>
>>> I've done it a few time and it isn't all that difficult: copy a similar
>>> >Is there no common "make an SMF service" helper out there yet? My searches did not yield fruit.
>>> >It would be really surprising if there still isnt one yet, after 7 years of SMF
>>>
>>> service XML file and change the name, dependencies, etc.
>>> Import them (svcadm restart manifest-import)
>>
>> that is semantically almost identical to,
>>
>> "It's not that hard to customise this program. Extract the source code, change the #defines in the right places, and type 'make'"
>>
>> That also falls in the realm of "not that difficult".
>> But it's not a very good 'customer interface' either.
>>
>>> (Clearly, this should be made simpler)
>>
>> I think this is the important take-away from this :) It's long past time that sun^H^H^Horacle provided a decent end-user tool.
>>
>> "Edit the xml file" is NOT "a decent end-user tool".
>> Especially one as convoluted as an SMF manifest.
>>
>> Take a look at the user-level questions that
>> http://code.google.com/p/manifold/
>> asks when you run it.
>>
>> Manifests could have an initial section of a few lines that hold those values, at the top. Rather than having to go through the entire file, "parsing" the xml in your head.
>> If xml is sooo simple to parse, why is libxml2.so 2 megabytes! :P
>> It's not nice to force humans to do something, that takes a machine many many lines of code to do.
>>
>> ugh.
>>
>> Personally, I would prefer a tool even simpler than the "manifold" tool. Something as simple as the redhat "chkconfig" tool.
>> That is to say, default to handling the 95% of most sysadmin's needs, with a minimum of unneccessary questions.
>>
>> It's almost always just a simple case of, "hey, I want this demon started at boot time, here's the start/stop script".
>>
>> Oracle is not meeting customer needs well in this area.
>
>I think it's safe to say solaris is dead this point.
>
>good job oracle.
>
>
So, what did they pay all that money for? Just patents?
(Against google, anyway, they seem to have lost their
gamble within the last week or two, getting probably
awarded 1% or less of what they expected. Or did
I read the newspaper wrong?)
If solaris "is dead at this point", then what are they going to do
for an OS? Linux on x86? I thought Suns were supposed to be
so much better architected for running Oracle databases?
Someone please comment on this "solaris is dead at this point"
statement.
THANKS!
David
Ian Collins
May 13, 2012, 10:45:16 PM5/13/12
to
Oracle may be doing a good job of driving away customers, but the genie
is out of the bottle and OpenSolaris based offerings are doing quite
well filling the void.
--
Ian Collins
Ian Collins
May 13, 2012, 10:47:05 PM5/13/12
to
--
Ian Collins
Ian Collins
May 13, 2012, 10:51:06 PM5/13/12
to
On 05/14/12 02:17 PM, David Combs wrote:
> In article<a0lnqp...@mid.individual.net>,
> Ian Collins<ian-...@hotmail.com> wrote:
>> On 05/ 6/12 08:54 AM, David Combs wrote:
>>>
>>> OK -- about Solaris seemingly moving away from ASCII
>>> config files:
>>>
>>> WHO is it who decides on such things?
>>
>> Oracle, who else?
>>
>>> Maybe some of the more knowledgeable in this group, say,
>>> could work to become outside advisors to or members
>>> of whatever groups decide this stuff.
>>>
>>> Likely the MBA-types will insist on GUI-type interfaces,
>>> a la Microsoft.
>>
>> No, if you look at what they are promoting, the centralised ops centre
>> is the path they are following.
>>
>>> BUT, even so, maybe at very low cost they could make available
>>> a parallel text-based interface.
>>
>> The intent appears to be an interface that is easy to script (and manage
>>from a central application). Text files are a pain to manage from scripts.
>
> In article<a0lnqp...@mid.individual.net>,
> Ian Collins<ian-...@hotmail.com> wrote:
>> On 05/ 6/12 08:54 AM, David Combs wrote:
>>>
>>> OK -- about Solaris seemingly moving away from ASCII
>>> config files:
>>>
>>> WHO is it who decides on such things?
>>
>> Oracle, who else?
>>
>>> Maybe some of the more knowledgeable in this group, say,
>>> could work to become outside advisors to or members
>>> of whatever groups decide this stuff.
>>>
>>> Likely the MBA-types will insist on GUI-type interfaces,
>>> a la Microsoft.
>>
>> No, if you look at what they are promoting, the centralised ops centre
>> is the path they are following.
>>
>>> BUT, even so, maybe at very low cost they could make available
>>> a parallel text-based interface.
>>
>> The intent appears to be an interface that is easy to script (and manage
>>from a central application). Text files are a pain to manage from scripts.
>
> Wow! Looks like I need some re-education!
>
> Updating multiple machines is something I know nothing about, being
> as I have only one Sun (sparc) up and running.
>
> Please, what are some of these non-text-file-based facilities that
> you and others here use to update hundreds of machines?
If I was looking after hundreds of machines,I would have a bigger house...
>
> Updating multiple machines is something I know nothing about, being
> as I have only one Sun (sparc) up and running.
>
> Please, what are some of these non-text-file-based facilities that
> you and others here use to update hundreds of machines?
> And suppose some of them are to have slight modifications to their
> setup? Like if its name has a "foo-" prefix, do this, else do that.
>
> Seems like an inherently programming-like problem, like where you
> have to tell the facility to do something it might never have
> done before, and probably is not already there as a built-in choice.
> Needs some kind of script, I'd think.
>
> So, how do these modern (gui-based?) facilities you guys use handle
> such requirements?
--
Ian Collins
Chris Ridd
May 14, 2012, 1:18:50 AM5/14/12
to
free "Less Known Solaris Features" (LKSF) book:
<http://www.c0t0d0s0.org/pages/lksfbook.html>
--
Chris
0 new messages