cancel
Showing results for 
Search instead for 
Did you mean: 

VCS cluster doesn't failover when NFS hangs / Network interface is down.

NSMC_Unix
Level 2

Hi,

I'm having an issue that whenever a NFS hangs (comes with message NFS not responding..) or whenever the Network interface goes down, the VCS cluster doesn't failover to its partner. Instead, it keeps trying on the same server and ultimately the resource group fails on both the nodes.

Can someone please help me out of this please? I would rather that the VCS fails over on its own or any other possible workaround to get it switchover manually, in case this scenario happens again. Is there anyway to achieve this?


Thanks & Regards,
Ajit

8 REPLIES 8

RiaanBadenhorst
Moderator
Moderator
Partner    VIP    Accredited Certified

post the engineA and nfs logs.

Attaching the logs.

Zahid_Haseeb
Moderator
Moderator
Partner    VIP    Accredited

under Service Group what resource type you used for NIC and NFS ?

Hi,

We are using Named & Namedpc for NIC and NFS.

Thanks,

Ajit

Zahid_Haseeb
Moderator
Moderator
Partner    VIP    Accredited

I think you need to o below:

- For Network card, you need to make NIC resource

- For NFS, you need to make NFSshare resource

 

Let me search in case I get a detailed technote

Zahid_Haseeb
Moderator
Moderator
Partner    VIP    Accredited

Page 255 Configuring NFS service groups

https://sort.symantec.com/public/documents/sfha/6.0/linux/productguides/pdf/vcs_admin_60_lin.pdf

 

I hope this will help you out

Thanks for the document. However, I have already got those resources configured. (I mean NIC resource exists on the NFS service service group). The problem is that even when its configured like this, the switchover doesn't happen if the NIC is down.

gnic_p         
gnic_p         
named_ip       
named_ip       
named_ipmp     
named_ipmp     
named_mount    
named_mount    
named_proc     
named_proc     
namedpc_ip     
namedpc_ip     
namedpc_ipmp   
namedpc_ipmp   
namedpc_mount  
namedpc_mount  
namedpc_proc   
namedpc_proc   
proxy-ipmp     
proxy-ipmp    

Eg:

#Resource    Attribute        System     Value
namedpc_ip   Group            global     namedpc
namedpc_ip   Type             global     IPMultiNICB

 

Thanks,
Ajit

frankgfan
Moderator
Moderator
   VIP   

NFS resource hung issue is a very popula issue and most of the times, these kind of issues are not VCS related.  If NFS is hanging (becoems not responsive, the issue is in the kernel), BVS NFS agent would not be able to offline/clean the hanging resource therefore unable to failover.  Likely a system reboot is needed.

Are you on VCS 5.0/5.1 or later?

If the issue occurs on a regula bases, patchj up OS and upgrade VCS is highly recommended.

If you are already on the latest OS/VCS patch level and still experience this issue, capture an application core dump (if its on Solaris) or just collect a live system call.  Log a support case with Veritas to tgroubleshoot the issue further.  To find the issue RCA, you need DataCollector (VRTSexplorer) report as well as an application or system core dump