Error:Probe is not completed on VCS nodes

[Anish]

Hi All,

I am new to VCS.

I have formed two node cluster on VMware server and trying to configure simple fail-over group of apache and Solaris zones (different)

Both Apache and Solaris Zones are on separate shared storage (vxvm disk group) as follows :

bash-3.00# vxdisk -o alldg  s list
DEVICE       TYPE            DISK         GROUP        STATUS
aluadisk0_0  auto:none       -            -            online invalid
c1t0d0s2     auto:none       -            -            online invalid
c3t0d0s2     auto:cdsdisk    -            (nfsdg)      online clone_disk ------ Currently DGs are deported
c3t1d0s2     auto:cdsdisk    -            (nfsdg)      online clone_disk
c3t2d0s2     auto:cdsdisk    -            (zonedg)     online clone_disk

main.cf file is attached with this FYR.

Regarding Apache Group:

When I start VCS on both nodes with hastart command

Status of ApacheG1 is unknown on both nodes as it is unbale to recognise the path /web-app (where apache config,statrup script available ) on shared storage. Due to this probe is not completed on both nodes.

So when i manualy import DG atleast first time on both nodes (one by one) probe operation completed sucessfully and then onwards i can switch Service group anywhere. But again when i stop VCS and start again it gives error probe not completed.

So if i want to configure fail-over apache SG do i need to mount DG and filesystem on both nodes?

Is this a best practice (mounting Filesystem on both nodes ) since it can accidently currupt by any user?

Or is there any other workaround ?

Regarding Zone Group :

Here also I am getting same error probe not completed.

I am able to start zone on both VCS nodes manually, but not able start using VCS.

For this I have created shared storage available to both VCS node : zonedg

First I mount Zonefilesyste :/test-zone and configured complete zone on single node and complete the installation procedure.

Then copied /etc/zone/test-zone/xml and index file to other VCS node.

I am able to start zones on both VCS nodes one by one manually but VCS is not able to start.

Pls suggest any workaround.

Thanks for your support.

Regards,

Anish

[Eric_Hennessey1 · Accepted solution]

Something's not quite right here...

When you started the cluster, VCS should have automatically imported the DG and brought all the rest of the resources online. Also, you mention that you're manually importing the DG on both nodes. Unless it's a CFS cluster shared diskgroup, how are you doing that? Or are you importing then deporting on one before importing on the other?

In any event, I'd suggest tailing the VCS engine log while the cluster is starting up to see just what's happening.

[mikebounds]

You cannot start the zone on both systems in VCS as it is in a failover group - if you want to do this then you need to put zone in a parallel group.  This is also when you can't import the diskgroup on both nodes, but if you put diskgroup in parallel service group then you need to use CVM&CFS as Eric says so you do not corrupt your data.

Is Appache running in the global zone or a non-global (local) zone?  If it is running in the local zone, then this is your problem, but if it is running in the global zone, then it looks like it is setup like example in bundled agents guide where binaries and conf are on shared storage so it should work. 

Mike