Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Hi,
I am trying to set up a dual-primary pair. As I had to learn that XFS
wasn't able to be run in dual-primary mode I killed one side of the DRBD
pair and set it up with GFS2. I followed the steps on
http://www.piemontewireless.net/Storage_on_Cluster_DRBD_and_GFS2 . After
setting it up I moved all the data from the XFS node to the GFS2 node,
verified that it's working, killed the XFS data, put GFS2 onto that node
and synced back.
My current status is that both nodes are in sync running "primary".
However, I am not able to mount /dev/drbd0 on node 2 ("xen2", the node
that was just synced). After executing the mount command it stalls until
I cancel with Ctrl+C.
Possibly the following outputs from node 2 help you helping me...
root at xen2:~# drbd-overview
0:drbd-sr1 Connected Primary/Primary UpToDate/UpToDate C r-----
root at xen2:~# tail /var/log/syslog:
[...]
Aug 9 11:59:07 xen2 gfs_controld[1960]: daemon cpg_join error retrying
Aug 9 11:59:09 xen2 fenced[1919]: daemon cpg_join error retrying
Aug 9 11:59:09 xen2 dlm_controld[1944]: daemon cpg_join error retrying
[...]
root at xen2:~# cman_tool nodes
Node Sts Inc Joined Name
1 M 104 2011-08-09 11:11:56 xen1
2 M 100 2011-08-09 11:11:56 xen2
When googling for the error I found that manual fencing on the properly
working node would help. However, this can cause file system damage - I
am a bit curious if I should really do that...
root at xen1:~# fence_ack_manual -n xen2
About to override fencing for xen2.
Improper use of this command can cause severe file system damage.
Continue [NO/absolutely]? NO
Aborted.
Any ideas?
CU,
Mészi.