5 Replies Latest reply on Feb 24, 2014 10:47 AM by Juergens-Oracle

Solaris Cluster 3.3u2 configuration - Solaris Ldoms

M10vir Feb 15, 2014 7:54 PM

Hi,

I've 2 sun blade T5-1B and I create 4 ldoms on each blade, also I install the Solaris Cluster 3.3u2 and its all done successfully, I did several time resource fail-over with nodes, everything seems working perfect for me as normal behavior of expected but I've an issue to test public transport testing, hereby I remove the both physical public transport cable form node1 but its didn't fail-over on node2, little confuse as its should be fail-over to node2, but not. Please let me know if where missed something to get proper fail-over?

Please advise.

regards,

1. Re: Solaris Cluster 3.3u2 configuration - Solaris Ldoms

Juergens-Oracle Feb 17, 2014 3:23 PM (in response to M10vir)

Hi M10vir,

this means you are running SC3.3u2 in guest domains and pulling the cables from the primary domain?

If so, this would mean that the IPMP group used in the guest domain need to get a link down that Solaris Cluster do a failover when logical hostname resources are configured. How are the network interfaces configured and what is the status of the network interfaces when the cables are removed? Furthermore does the domains recognize that the network is gone and the relevant messages are in /var/adm/messages?

Thanks,
Juergen
Like Show 0 Likes(0)

Actions
2. Re: Solaris Cluster 3.3u2 configuration - Solaris Ldoms

M10vir Feb 18, 2014 7:48 PM (in response to Juergens-Oracle)

Thanks Juergen,
Yes, you're right, but general behavior of Cluster it shouldn't be fail-over while we do forcefully, In general we've dedicated network interface like bge0 & bge1 for public and configured the link based IPMP.
When I unplug the cable in control domain its showing DOWN but in guest domain its showing UP

please advise

Thanks
Like Show 1 Likes(1)

Actions
3. Re: Solaris Cluster 3.3u2 configuration - Solaris Ldoms

Juergens-Oracle Feb 19, 2014 2:57 PM (in response to M10vir)

Hi M10vir,

ok, this means the issue seems to be in the vsw of the Oracle VM Server (LDom) software because the control domain does not pass the down information to the guest domain. Please can you open a SR for LDom software with your configuration details (explorer outputs) that we can identify if this is a bug.
Maybe you can also try probe-based IPMP to see if the problem is persistent?

Thanks,
Juergen
Like Show 0 Likes(0)

Actions
4. Re: Solaris Cluster 3.3u2 configuration - Solaris Ldoms

M10vir Feb 22, 2014 5:07 PM (in response to Juergens-Oracle)

Sure, I'll try with probe-based IPMP first also will try to get the explorer just point me where to upload it or you need some pacific output?

Thanks,
Like Show 0 Likes(0)

Actions
5. Re: Solaris Cluster 3.3u2 configuration - Solaris Ldoms

Juergens-Oracle Feb 24, 2014 10:47 AM (in response to M10vir)

Hi M10vir,
before you can upload explorer you need to open a SR for LDom Software. Within the SR the SR owner will let you know what additional information we need to identify the issue.

Thanks,
Juergen
Like Show 0 Likes(0)

Actions

Go to original post

Join the world’s largest interactive community dedicated to Oracle technologies.