Forum Discussion

symsonu's avatar
symsonu
Level 6
12 years ago

solution needed for vxfen issue

 

there is a two node cluster and we split two node cluster for upgrade. The isolated node is not coming up as vxfen is not starting



/02/01 11:52:05 VCS CRITICAL V-16-1-10037 VxFEN driver not configured. Retrying...
2013/02/01 11:52:20 VCS CRITICAL V-16-1-10037 VxFEN driver not configured. Retrying...
2013/02/01 11:52:35 VCS CRITICAL V-16-1-10037 VxFEN driver not configured. Retrying...
2013/02/01 11:52:50 VCS CRITICAL V-16-1-10037 VxFEN driver not configured. Retrying...
2013/02/01 11:53:05 VCS CRITICAL V-16-1-10037 VxFEN driver not configured. Retrying...
2013/02/01 11:53:20 VCS CRITICAL V-16-1-10031 VxFEN driver not configured. VCS Stopping. Manually restart VCS after configuring fencing
^C


IOFENCING configuration seems okay on node 2 as /etc/vxfentab has the entry for  co-ordinator disks and /etc/vxfendg has diskgroup entry vxfendg2 and these disks and diskgroup are visible too.

DEVICE       TYPE            DISK         GROUP        STATUS
c0t5006016047201339d0s2 auto:sliced     -            (ossdg)      online
c0t5006016047201339d1s2 auto:sliced     -            (sybasedg)   online
c0t5006016047201339d2s2 auto:sliced     -            (vxfendg1)   online
c0t5006016047201339d3s2 auto:sliced     -            (vxfendg1)   online
c0t5006016047201339d4s2 auto:sliced     -            (vxfendg1)   online
c0t5006016047201339d5s2 auto:sliced     -            (vxfendg2)   online
c0t5006016047201339d6s2 auto:sliced     -            (vxfendg2)   online
c0t5006016047201339d7s2 auto:sliced     -            -            online
c2t0d0s2     auto:SVM        -            -            SVM
c2t1d0s2     auto:SVM        -            -            SVM

On checking vxfen.log

 

nvoked S97vxfen. Starting
Fri Feb  1 11:50:37 CET 2013 starting vxfen.. 
Fri Feb  1 11:50:37 CET 2013 calling start_fun
Fri Feb  1 11:50:38 CET 2013 found vxfenmode file
Fri Feb  1 11:50:38 CET 2013 calling generate_vxfentab
Fri Feb  1 11:50:38 CET 2013 checking for /etc/vxfendg
Fri Feb  1 11:50:38 CET 2013 found  /etc/vxfendg.
Fri Feb  1 11:50:38 CET 2013 calling generate_disklist
Fri Feb  1 11:50:38 CET 2013 Starting vxfen.. Done. 
Fri Feb  1 11:50:38 CET 2013 starting in vxfen-startup
Fri Feb  1 11:50:38 CET 2013 calling regular vxfenconfig
Fri Feb  1 11:50:38 CET 2013 return value from above operation is 1
Fri Feb  1 11:50:38 CET 2013
output was VXFEN vxfenconfig ERROR V-11-2-1003 At least three coordinator disks must be defined
Log Buffer: 0xfffffffff4041090


refadm2-oss1{root} # cat /etc/vxfendg
vxfendg2

and there are below mentioned two disks in vxfendg2

c0t5006016047201339d5s2 auto:sliced     -            (vxfendg2)   online
c0t5006016047201339d6s2 auto:sliced     -            (vxfendg2)   online

 

is it due to two  disks in coordinator diskgroup? Is it a known issue ?

 


 

  • Try: 

    vxdg -k -g <group> adddisk <disk>=<device> 

    vxdg -k -g vxfendg2 adddisk c0t5006016047201339d7s2=c0t5006016047201339d7s2

    (We don't normally use Device name for Disk name, but it seems this is what you've done for other disks in vxfendg2).

  • Try: 

    vxdg -k -g <group> adddisk <disk>=<device> 

    vxdg -k -g vxfendg2 adddisk c0t5006016047201339d7s2=c0t5006016047201339d7s2

    (We don't normally use Device name for Disk name, but it seems this is what you've done for other disks in vxfendg2).

  • You could: 1) disable fencing (/opt/VRTS/install/installvcs -fencing) and then enable fencing. 2) run vxfenswap to change fencing devices (https://sort.symantec.com/public/documents/sfha/6.0.1/linux/productguides/html/sfsybasece_admin/ch02s03s05.htm)
  • Hello,

     

    what does -k  option specifies in  below coomand

     

    vxdg -k -g vxfendg2 adddisk c0t5006016047201339d7s2=c0t5006016047201339d7s2

     

    I mean the use of -k option please.

     

     

  • re: vxdg -k option - from manpage:

    https://sort.symantec.com/public/documents/sfha/5.1sp1/solaris/manualpages/html/manpages/volume_manager/html/man1m/vxdg.1m.html

    ----------
    KEYWORDS

    adddisk

    [...]

    If the -k option is specified, the disk media name must represent a disk media record that was previously dissociated from its disk access record with -k rmdisk; otherwise, a new disk media record is created to represent the disk. With the -k option, plexes requiring recovery are flagged as stale.
    ----------

    minor amendment to Marianne's command as the other dmnames don't use s2:

    # vxdg -g vxfendg2 -k adddisk c0t5006016047201339d7=c0t5006016047201339d7s2

    If the adddisk doesn't work, please provide the output from:

    # vxdisk list c0t5006016047201339d7s2

  • I have added the disk with below command

     

    vxdg -g vxfendg2 -o coordinator -kf adddisk c0t5006016C47201339d7=c0t5006016047201339d7s2.

     

    a slight modification needed to add -o coordinator as its a coordinator diskgroup

    Thanks all for the excellent support

     

     

  • Please mark Marianne's post as solution since she pointed out the command.