[Rocks-Discuss] [zsm] rocks rock5.4.3.x86_64 vmware esx5

67 views
Skip to first unread message

Gerhard W. Recher (privat)

unread,
Mar 27, 2012, 8:08:10 AM3/27/12
to npaci-rocks...@sdsc.edu
hi @all

I have a serious problem,

I installed rocks frontend with all avail rolls from dvd iso

after install I can not ping to eth1

Vmware tools are not installed

e1000driver is loaded, i See arp entry of default gateway, but no ping
to outside is possible.

stopping iptables makes no difference.

I installed with the same configuration a ubuntu 11.10 and have no
problems with networking at all, alo without vmware tools installed


so I'm stucked, any hints how to test rocks in vm ware esx 5 environment ?

-- Gerhard


-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/attachments/20120327/314f1c9b/attachment.html
-------------- next part --------------
A non-text attachment was scrubbed...
Name: dafbegge.png
Type: image/png
Size: 219897 bytes
Desc: not available
Url : https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/attachments/20120327/314f1c9b/dafbegge.png

Daniele

unread,
Mar 27, 2012, 8:55:20 AM3/27/12
to Discussion of Rocks Clusters
I run rocks frontend (viper 5.4.3) on ESXi 5 without problem with e1000
driver or vmxnet.

eth0 -> private network for cluster nodes
eth1 -> "public" ip
default gateway on eth1

Daniele

Gerhard W. Recher (privat)

unread,
Mar 27, 2012, 9:37:28 AM3/27/12
to Discussion of Rocks Clusters
i now installed vmware tools, and reboot.

up to the point "starting xend:" i'm able to ssh into frontend, but
shortly later, connectifity vanished.

ping cluster1
PING cluster1.netpilot.net (62.67.240.41) 56(84) bytes of data.
64 bytes from cust-240-41.netpilot.net (62.67.240.41): icmp_seq=1 ttl=63
time=6.57 ms
64 bytes from cust-240-41.netpilot.net (62.67.240.41): icmp_seq=2 ttl=63
time=5.35 ms
64 bytes from cust-240-41.netpilot.net (62.67.240.41): icmp_seq=3 ttl=63
time=3.52 ms
64 bytes from cust-240-41.netpilot.net (62.67.240.41): icmp_seq=4 ttl=63
time=6.05 ms
^C64 bytes from cust-240-41.netpilot.net (62.67.240.41): icmp_seq=5
ttl=63 time=17.5 ms

--- cluster1.netpilot.net ping statistics ---
5 packets transmitted, 5 received, 0% packet loss, time 8153ms
rtt min/avg/max/mdev = 3.524/7.813/17.567/4.985 ms
dbserv:/var/www/clean-mx# slogin cluster1
The authenticity of host 'cluster1 (62.67.240.41)' can't be established.
RSA key fingerprint is a4:23:7d:19:2a:29:92:05:99:ed:15:dc:f5:5a:0e:f5.
Are you sure you want to continue connecting (yes/no)? yes
Warning: Permanently added 'cluster1,62.67.240.41' (RSA) to the list of
known hosts.
root@cluster1's password:
Last login: Tue Mar 27 11:48:10 2012
Rocks 5.4.3 (Viper)
Profile built 21:19 26-Mar-2012

Kickstarted 00:49 27-Mar-2012

[root@cluster1 ~]#
[root@cluster1 ~]# netstat -rn
Kernel IP routing table
Destination Gateway Genmask Flags MSS Window irtt
Iface
62.67.240.32 0.0.0.0 255.255.255.224 U 0 0 0
eth1
192.168.122.0 0.0.0.0 255.255.255.0 U 0 0 0
virbr0
10.1.0.0 0.0.0.0 255.255.0.0 U 0 0 0
eth0
169.254.0.0 0.0.0.0 255.255.0.0 U 0 0 0
eth0
0.0.0.0 62.67.240.33 0.0.0.0 UG 0 0 0
eth1
[root@cluster1 ~]#


network table inside vmware console look now (see attached pic ....
will be scrubbed form mailman...)


-- gerhard

URL: https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/attachments/20120327/426a3cb4/attachment.html

-------------- next part --------------
A non-text attachment was scrubbed...

Name: ibbbcfja.png
Type: image/png
Size: 83096 bytes
Desc: not available
Url : https://lists.sdsc.edu/pipermail/npaci-rocks-discussion/attachments/20120327/426a3cb4/ibbbcfja.png

Luca Clementi

unread,
Mar 28, 2012, 12:38:12 PM3/28/12
to Discussion of Rocks Clusters
Gerhard,
as of my yesterday post:

>>
>> and worst: after this action firewall command is not recognized..
>>
>>  rocks open host firewall localhost network=public protocol=tcp service=www
>> error - invalid rocks command "open"
>
> The proper command is
> rocks add firewall {category=index} [action=string] [chain=string]
> [network=string]
>        [output-network=string] [protocol=string] [rulename=string]
>        [rulesrc=string] [service=string]
>
> There is an error in the documentation.
>
> You can use
> rocks add firewal help
> to get more info.
>
>


On Tue, Mar 27, 2012 at 11:40 AM, Gerhard W. Recher (privat)
<ger...@recher.de> wrote:
> Luca,
>
> thx for reply... I started from scratch installing a frontend with all
> rolls but excluded xen.
>
> i still have on a fresh system this "open" problem
>
>  rocks open host firewall localhost network=public protocol=tcp service=www
> error - invalid rocks command "open"
> [root@cluster1 ~]#
>
> so i logged in to gui and made port 80 an 443 working with fireall settings in the gui.
>
>
> this is apparently a bad idea, because ganglia stopped working
>
> I have no idea why a fresh install has this behavior, from my point of view I can not do wrong steps by installing a frontend by following the documentation.
>
> any advice ?
>
>
> -- Gerhard
>
>
>
> Am 27.03.2012 18:33, schrieb Luca Clementi:
>> Dear Gerhard,
>> Xen is a virtualization system and is not supposed to be run under
>> another virtualization system.
>>
>> It might work (using para-virtualization), but its a source of trouble
>> and there are a lot of tips and tricks necessary to make it work.
>> I was able to run XEN under KVM but it took me a while to get it
>> working properly
>> (http://rwmj.wordpress.com/2010/10/07/tip-run-xen-as-a-kvm-guest/)
>>
>> But generally I would strongly recommend to use xen roll _only_ on bare metal.
>>
>> Sincerely,
>> Luca
>>
>>
>>
>> On Tue, Mar 27, 2012 at 7:30 AM, Gerhard W. Recher (privat)
>> <ger...@recher.de> wrote:
>>> I found a partial solution.
>>>
>>> i disabled xend an xendomains , so network is now ok.
>>>
>>> but why does xen affect networking in such a worse manor ?
>>>
>>>
>>> and worst: after this action firewall command is not recognized..
>>>
>>>  rocks open host firewall localhost network=public protocol=tcp service=www
>>> error - invalid rocks command "open"
>>> [root@cluster1 ~]#
>>>
>>>
>>> i'm really stuck...
>>>
>>>
>>> -- Gerhard

Reply all
Reply to author
Forward
0 new messages