Well? Well? Anything?

Jeff Wolfe

unread,

Jan 4, 2000, 3:00:00 AM1/4/00

to

In article <94695561...@sidehack.sat.gweep.net>,
MegaZone <mega...@news.gweep.net> wrote:
>wo...@ems.psu.edu (Jeff Wolfe) shaped the electrons to say:
>>Even my trusty 8 yr old Annex-III is still chugging along. It's logging dates
>>as "1000101" but we can work with time_t problems.. :)
>
>Shit, Xylogics flashbacks...
>
>What rev of code are you running on that beast?

annex: stat
S/W Version: ANNEX3-UX R7.1 Build #10: Mon Dec 20 06:06:29 EST 1993
ROM Rev: . H/W Type: Annex3 16s/1p.
H/W Rev: 8.0. Serial Number: 152110.
Uptime: 119 days 2 hours 59 minsDate: Tue Jan 4 13:16:14 2000 EST
Boot from: 0.0.128.17 Image: yy!yyyyyyyyyyyyyyyyyyyyy
Subnet mask: 255.255.255.0 Broadcast addr: 128.118.
Inet addr: 128.118. Ethernet addr: 8a-00-d1-84-ec-11
Default domain: ems.psu.edu
Loading:
CPU = 5% procs active/max/limit = 32/37/240
rescheds = 0/593 switches = 42/13467176 activates = 43/13472822
Callouts:
max=227 free=215 minimum free=212
Mbufs:
total=1799 free=1346 minimum free=1009 denied=0
Serial Ports:
Total bytes: rcv'd=2044991 xmt'd=706133325
Errors: parity=0 framing=3 fifo overruns=0
Parallel Ports:
Total bytes: xmt'd=0
Memory:
total=1572864 avail=897560 free=221712 min free=182992 fails=0
annex:

You'll note the NVRAM is hosed.. I can still boot it, but I have to
reset the IP address in the PROM every time it's power cycled.

go figure..

-Jeff

--

Jeff Wolfe College of Earth and Mineral Science - Penn State

Peter da Silva

unread,

Jan 4, 2000, 3:00:00 AM1/4/00

to

In article <84rggu$iei$1...@laurasia.ems.psu.edu>,
Jeff Wolfe <wo...@ems.psu.edu> wrote:
>So far, everyone has been grudgingly willing to switch. Except this
>one guy.. He said: "Oh, it puts 100 in for the year? Well I'll keep
>using it and think about switching if it bothers me." we said "you
>better find somthing else by the end of the week."

My response to that would be to install my own copy of Elm 2.4-ME-PL61
in my home directory.

>Even my trusty 8 yr old Annex-III is still chugging along. It's logging dates
>as "1000101" but we can work with time_t problems.. :)

That's a (struct tm) problem, silly.

--
This is The Reverend Peter da Silva's Boring Sig File - there are no references
to Wolves, Kibo, Discordianism, or The Church of the Subgenius in this document

Executive Vice President, Corporate Communications, Entropy Gradient Reversals.

David Iain Greig

unread,

Jan 4, 2000, 3:00:00 AM1/4/00

to

Peter Radcliffe <26$10$f3i...@pir.net> wrote:
>David Iain Greig <gr...@ediacara.org> probably said:
>>The top two answers to 'how many days in the last month did you drink'
>>at LISA in Seattle were:
>>
>> 0 and 1.
>
>That was just fucked up.

I dunno, I think I've drunk 2 or 3 days in the last 30. That would be
Xmas day (icewine), New Years (bubbly at 1:30 after the sturm und drang
failed to occur) and once in a bar I had a beer. Take out the holidays and
you get ... 1.

I don't see the need to get wasted. I did it when I was in my early 20's,
but nowadays I like my mornings to be coherent and non-hungover. I just
don't need to get drunk to have a good time. Maybe I have a low-stress
job; UNIX admin is , after all, pretty damn easy most of the time.

YMMV.

--D.

Jeff Wolfe

unread,

Jan 4, 2000, 3:00:00 AM1/4/00

to

In article <84tg19$6m5$1...@citadel.in.taronga.com>,

Peter da Silva <pe...@taronga.com> wrote:
>In article <84rggu$iei$1...@laurasia.ems.psu.edu>,
>Jeff Wolfe <wo...@ems.psu.edu> wrote:
>>So far, everyone has been grudgingly willing to switch. Except this
>>one guy.. He said: "Oh, it puts 100 in for the year? Well I'll keep
>>using it and think about switching if it bothers me." we said "you
>>better find somthing else by the end of the week."
>
>My response to that would be to install my own copy of Elm 2.4-ME-PL61
>in my home directory.

Heh.. :)

>>Even my trusty 8 yr old Annex-III is still chugging along. It's logging dates
>>as "1000101" but we can work with time_t problems.. :)
>
>That's a (struct tm) problem, silly.

Yup.. Seen lots of those.. Guess Nortel could open-source the Annex-UX
code so I could fix it, but I'm just a PRI away from installing a
replacement, so I'm not particularly worried about it. The problem
seems especially prevalent with perl CGI scripts..

unread,

Jan 4, 2000, 3:00:00 AM1/4/00

to

David Iain Greig <gr...@ediacara.org> writes:

> Most of the answers to the quiz didn't suprise me; they were 'corporate
> UNIX admin' answers... using Netbackup and ADSM and the like versus
> 'tar'. Evi Nemeth was sitting next to me, and she couldn't believe the
> answers. I could; they were all normal for peoplw who work in sizable
> datacenters with budgets... not necessarily the sort of people who
> build USENIX...

Hm... we're using ADSM. Does that make us corporate? :) Expensive SoB,
but it does the job as advertised.

--
Russ Allbery (r...@stanford.edu) <URL:http://www.eyrie.org/~eagle/>

Bryan C. Andregg

unread,

Jan 5, 2000, 3:00:00 AM1/5/00

to

On 4 Jan 2000 18:19:37 GMT, Jeff Wolfe <wo...@ems.psu.edu> wrote:
>annex: stat

>S/W Version: ANNEX3-UX R7.1 Build #10: Mon Dec 20 06:06:29 EST 1993

>ROM Rev: . H/W Type: Annex3 16s/1p.
>H/W Rev: 8.0. Serial Number: 152110.

NNNNNNNNNNOOOOOOOOOOOOOOOOO!!!!!!!!!!

Get out of my head!

--
Bryan C. Andregg * <band...@redhat.com> * Red Hat, Inc.

1024/625FA2C5 F5 F3 DC 2E 8E AF 26 B0 2C 31 78 C2 6C FB 02 77
1024/0x46E7A8A2 46EB 61B1 71BD 2960 723C 38B6 21E4 23CC 46E7 A8A2

Jeff Wolfe

unread,

Jan 5, 2000, 3:00:00 AM1/5/00

to

In article <ylk8lp4...@windlord.stanford.edu>,

Russ Allbery <r...@stanford.edu> wrote:
>David Iain Greig <gr...@ediacara.org> writes:
>
>> Most of the answers to the quiz didn't suprise me; they were 'corporate
>> UNIX admin' answers... using Netbackup and ADSM and the like versus
>> 'tar'. Evi Nemeth was sitting next to me, and she couldn't believe the
>> answers. I could; they were all normal for peoplw who work in sizable
>> datacenters with budgets... not necessarily the sort of people who
>> build USENIX...
>
>Hm... we're using ADSM. Does that make us corporate? :) Expensive SoB,
>but it does the job as advertised.

We're using AMANDA for the *nix world and BackupExec for the Other Stuff...
We're about to write a shim to get AMANDA to work with StorageTek's ACSLS so
we can use the shiny new 7000 cell Powderhorn silo we just got.. We're also
going to be experimenting with ADSM (which was free under our IBM contract
until they sold adsm to Tivoli.) for HSM and backing up our SP.

Biggest problem I have with most of the commercial apps is the
non-standard nature of the files on the backup media.. At least with
Amanda, if the whole world dies and all I have are my tapes, I know I
can figure out what's what and restore with nothing more than a
working OS. I don't want to have to keep track of licenses, "disaster
recovery disks" and other junk so I can restore my backups in an
emergency.

Jeff Wolfe

unread,

Jan 5, 2000, 3:00:00 AM1/5/00

to

In article <94702674...@sidehack.sat.gweep.net>,

MegaZone <mega...@news.gweep.net> wrote:
>wo...@ems.psu.edu (Jeff Wolfe) shaped the electrons to say:

>>S/W Version: ANNEX3-UX R7.1 Build #10: Mon Dec 20 06:06:29 EST 1993
>

>Wow, 7.1 :-) I figured R8.0.8 or X8.0.24 or so. 7.1 is fairly minimal,
>but hey, it was one of the more stable releases, so whatever works. :-)

It's not upgradeable past 7.1, according to Bay. Somthing about
memory.

Strange

unread,

Jan 6, 2000, 3:00:00 AM1/6/00

to

Abby Franquemont <ab...@ucan.foad.org> wrote:
> I also heard that everyone said they used Veritas -- which, as was
> pointed out by the fellow relaying this to me, just doesn't add up
> with the nobody drinking thing.

As their CTO informed us under some pressure, Veritas uses
challenge-response authentication.

After some hard questioning, it was revealed that the challenge is always,
"Hi, I am a Veritas server!" and the response is always, "Hi, I'm a
Veritas client!"

Which is to say I think I understand what you're saying.

-M

--
Michael Brian Scher (MS683/MS3213)| Anthropologist, Attorney, Policy Analyst
Mainlining Internet Connectivity for Fun and Profit
str...@netural.com li...@foad.org str...@cultural.com str...@ispfh.org
Give me a compiler and a box to run it, and I can move the mail.

David Iain Greig

unread,

Jan 7, 2000, 3:00:00 AM1/7/00

to

nic...@esperi.org <nic...@esperi.org> wrote:
>
>One of the reasons I think Veritas NetBackup sucks somewhat less than
>other options is that it uses, mostly, TAR format for its tapes. So you can,
>given an appropriately-prepared tape (ie, it has to have all the files
>for the machine on it, not just an incremental) boot it from (cd/floppy/net)
>and use the OS utilities to restore it (if you can prove me wrong, please do).

No, NetBackup uses tar. There's a label written to each tape as well
to allow the machine to tell what tape you popped in (unless you're
using a silo/robot/whathaveyou, in which case it ought to know, and
the reader can scan the label anyhow).

Apart from that, if you have multiplexed backups on the same tape, then
untarrring it on a naive (non NetBackup-managed) tape drive would be,
in all probability, a nightmare. You can multiplex several concurrent
backups onto the same tape in order to improve throughput. So you have
non-sequential blocks of data on the tape.

Our DRP plan (as executed in test) is:
1) go to recovery site (contracted to provide hardware)
2) wait for tapes to arrive
3) - install Sun NetBackup Master on bare metal by CD
- configure designated RS/6000 servers with mksysbs
4) start restore of tape catalogues to horking big disk array
5) start rebuilding volume groups (empty) on EMC/JBOD/RAID whatever
6) when filesystems recreated, start restoring data from NetBackup
7) sit back and watch a fractional-terabyte database restore go

Sun's bare-metal restore is shockingly primitive. With mksysb or Ignite
tapes, I pop a tape in, reboot, and eventually, with little or no
admin input, I have a bootable system that simply lacks data volumes. It
scales quite nicely... I can pop 20 tapes in 20 machines by myself and
watch them all restore autonomously. Recreating masses of logical volumes
is a killer timewise. But you can script a lot of it ahead of time and
simply put your recovery scripts in rootvg to get caught by weekly mksysb.

Aside: if you use VxVM, and you have identical hardware, then if you make
a copy of vxprint -list -vmrs (or some such, see Veritas web site for
tech tip dated Nov 23) you can feed that to vxmake and recreate your volumes
bing bang boom. Found that our in the last few days... too late to be of use.

The new billing system on a cluster of V-class HP's uses EMC and SRDF[1]
to mirror the data drives to a hotsite downtown (datacenter in burbs),
with (I think) 2 T3's to handle the traffic. For some reason, while you can
boot off EMC disks under HP/UX. something about MC/ServiceGuard or fibre-
channel SCSI breaks that. Otherwise we'd be using SRDF to mirror the boot
drives (less paging spaces :P) and we'd have the billing system with
recovery on the order of a few minutes. Beats restoring 10 TB of data. :)

One problem with NB is you can't give an operator a limited function
off-the-shelf tool to check backups without them being able to seriously
pooch something. StorageTek offered to write us some stuff, if we'd pay,
though. We passed.

--D.

[1] Symmetrix Remote Data Facility, or something. Mirrors yer data to
a distant site over (insert fat pipe here).

Strange

unread,

Jan 7, 2000, 3:00:00 AM1/7/00

to

Strange has been having brain-off days, it seems. The vendor using *cough
cough* "challenge-response" authentication is Legato. Veritas actually
has a slightly customized version of OPIE compiled in (they call it
VOPIE).

-M