[Pacemaker] iscsi target mounting readonly on client

Joseph-Andre Guaragna joseph-andre at rdmo.com
Fri Apr 12 15:19:17 UTC 2013


Hi,

I just finish to install a cluster based on pacemaker for High
available ISCSI target (active/passive), everythinf work like a charm
swictching between nodes, fencing rebooting and disconnecting node.

the cluster runs the following properties ipaddr2, iscsi target, iscsi
lun and drbd.

But here comes the troubles as soon as I try too bench the iscsi and
move service accross nodes, the iscsi target get remounted as
read-only.

I try to tweak on the client /etc/iscsi/iscsid.conf
 with
node.conn[0].timeo.logout_timeout = 10
node.conn[0].timeo.noop_out_interval = 30
node.session.timeo.replacement_timeout = 480


I also modigy the ISCSI target provider /etc/iet/ietd.conf


DefaultTime2Wait        5
DefaultTime2Retain      60

On the client I get the followinfg error messages:


 sd 14:0:0:0: [sdk] Attached SCSI disk
[ 4146.898335] EXT4-fs (sdk1): recovery complete
[ 4146.900299] EXT4-fs (sdk1): mounted filesystem with ordered data
mode. Opts: (null)
[ 4173.975058] EXT4-fs error (device sdj1): ext4_put_super:819:
Couldn't clean up the journal
[ 4173.975065] EXT4-fs (sdj1): Remounting filesystem read-only
[ 4182.347630] EXT4-fs error (device sdi1): ext4_put_super:819:
Couldn't clean up the journal


I any of you have a way to avoird the disconnection and the read-only
remounting I would be glad.


Joseph




More information about the Pacemaker mailing list