cancel
Showing results for 
Search instead for 
Did you mean: 

could not eject node from disk

redtred
Level 2

Hello Guys,

 

Please help me to analize and make a solution to this problem. I have a cluster and i am testiong a failover from server1 to server2 then i encountered this problem in which the disk could not be ejected and leads into kernel panic. Our OS is Redhat5.5. Below is the procedure i made for the failover and the problem i encountered. Thank you in advance.

  1. Failover testing from server1 to server2.
  2. Procedure is to shutdown the interfaces on server1 eth5 and bond2 . eth0 is the hearbeat and bond2 is the secondary LLT.
  3. After shutdown of the interfaces the services VIPS and mountpoints should kick in to server2.
  4. The failover failed instead the server2 had a kernel panic. We did it twice same result server2 is panicking.
  5. Error code message is captured on server2: V-11-1-65 and V-11-1-20
  6. Last test we did again in which we shutdown the server1 interfaces eth0 and bond2.262 server1  itself went into panic mode.
  7. Same error code on server1: V-11-1-65 and V-11-1-20

 

2 REPLIES 2

Gaurav_S
Moderator
Moderator
   VIP    Certified

Would be of more help to paste exact error messages ...

Also, Do you use IOFencing ?

# rpm -aq  | egrep 'fen|gab|llt'

# grep -i usefence /etc/VRTSvcs/conf/config/main.cf

you are expecting that after shutting down the VIPS the mount resources should be failed over to other group ... are your VIPS marked as "critical" resources in service group ?

Refer VCS admin guide to know how to mark resources as critical .. refer page 143 on below link

https://sort.symantec.com/public/documents/vcs/5.1/linux/productguides/pdf/vcs_admin.pdf

Also, paste the messages before & during panic...

Gaurav

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

Please post engine_A log (in /var/VRTSvcs/log) as well as messages files on both nodes. Please post these files as attachments.