[Pacemaker] gfs2

Andrew Beekhof andrew at beekhof.net
Fri Aug 27 13:53:43 UTC 2010 

On Mon, Aug 9, 2010 at 6:43 PM, bunkertor <bunkertor at tiscali.it> wrote:
> hi to all!
> i have some problems with cluster and gfs2 mounting on iscsi devices.
> i followed this guide
> "http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf" with some changes
> depending on my envoironment; so the cluster should connect to iscsi portal,
> start dlm_controld(.pcmk) and gfs_controld(.pcmk) and mount gfs2 partitions
> on both nodes.
> but when cluster tries to mount gfs2 partition on both nodes, i got an error
> in corosync.log, like this:
>
> Aug 09 17:30:57 tromso lrmd: [7495]: info: RA output: (dlm:1:start:stderr)
> dlm_controld.pcmk: no process found
> ....
> Aug 09 17:30:58 tromso lrmd: [7495]: info: RA output:
> (gfs2FS:0:start:stderr) error mounting /dev/sda on /mnt/fs-cl00: No such
> file or directory

Perhaps /dev/sdaN rather than /dev/sda ?

>
> in dmesg:
> .....
> [14000.428842] GFS2: can't find protocol lock_dlm
>
> but dlm_controld.pcmk is up and running
> tromso:/etc/corosync # ps -eaf | grep pcmk
> root      7699     1  0 17:30 ?        00:00:00 dlm_controld.pcmk -q 0
> root      7704     1  0 17:30 ?        00:00:00 gfs_controld.pcmk -g 0

What does it say when you run the following:
      killall -0 dlm_controld.pcmk; echo $?


>
> and mount point exists
> tromso:/etc/corosync # ls -l /mnt/
> total 4
> drwxr-xr-x 2 root root 4096 Aug  9 11:41 fs-cl00
>
> in gfs_controld it logs this error aswell:
>
> Aug 09 17:36:01 gfs_controld mount_done: fscl not found
> Aug 09 17:36:01 gfs_controld do_leave: fscl not found

fscl ?

>
> iscsi, dlm and gfs2 control clones work fine.

well, not if its saying:
> Aug 09 17:30:57 tromso lrmd: [7495]: info: RA output: (dlm:1:start:stderr)
> dlm_controld.pcmk: no process found

unless thats from the check _before_ the daemon is started.

> when i try to mount partition manually i have this output: [where
> gfs2-cluster in clustername in corosync.conf and fscl is the name of 2
> journals]

Can you show us your corosync.conf?

>
> tromso:/etc/corosync # mount -t gfs2 /dev/sda /mnt/fs-cl00 -o
> locktable=gfs2-cluster:fscl,lockproto=lock_dlm -vv
> mount /dev/sda /mnt/fs-cl00
> parse_opts: opts = "rw,locktable=gfs2-cluster:fscl,lockproto=lock_dlm"
>  clear flag 1 for "rw", flags = 0
>  add extra locktable=gfs2-cluster:fscl
>  add extra lockproto=lock_dlm
> parse_opts: flags = 0
> parse_opts: extra = "locktable=gfs2-cluster:fscl,lockproto=lock_dlm"
> parse_opts: hostdata = ""
> parse_opts: lockproto = "lock_dlm"
> parse_opts: locktable = "gfs2-cluster:fscl"
> lock_dlm_join: hostdata: "hostdata=jid=0:id=382599105:first=1"
> mount(2) failed error -1 errno 2
> error mounting /dev/sda on /mnt/fs-cl00: No such file or directory
>
> and...
>
> tromso:/etc/corosync # gfs2_tool sb /dev/sda table
> current lock table name = "gfs2-cluster:fscl"
> tromso:/etc/corosync # gfs2_tool sb /dev/sda proto
> current lock protocol name = "lock_dlm"
>
> i cannot see where i make mistakes, so i'm here to ask help =)
>
> here my cluster configuration:
> ===============
> tromso:/etc/corosync # crm configure show
> node oslo
> node tromso
> primitive Cluster-IP ocf:heartbeat:IPaddr2 \
>        params ip="172.16.1.200" cidr_netmask="32" \
>        op monitor interval="30s"
> primitive dlm ocf:pacemaker:controld \
>        op monitor interval="120s"
> primitive gfs-control ocf:pacemaker:controld \
>        op monitor interval="120s"
> primitive gfs2FS ocf:heartbeat:Filesystem \
>        params device="/dev/sda" directory="/mnt/fs-cl00" fstype="gfs2"
> primitive iscsi ocf:heartbeat:iscsi \
>        params iscsiadm="/sbin/iscsiadm"
> target="iqn.2010-07.uk.co.bunkertor:storage.lvm02"
> portal="172.16.1.101:3260"
> clone dlm-clone dlm \
>        meta interleave="true"
> clone gfs-clone gfs-control \
>        meta interleave="true"
> clone gfs2FS-clone gfs2FS \
>        meta interleave="true"
> clone iscsi-clone iscsi \
>        meta interleave="true"
> colocation c_gfs2 inf: gfs2FS-clone ( gfs-clone dlm-clone ) iscsi-clone
> order o_gfs2 inf: iscsi-clone ( dlm-clone gfs-clone ) gfs2FS-clone
> property $id="cib-bootstrap-options" \
>        dc-version="1.0.9-74392a28b7f31d7ddc86689598bd23114f58978b" \
>        cluster-infrastructure="openais" \
>        expected-quorum-votes="2" \
>        stonith-enabled="false" \
>        no-quorum-policy="ignore" \
>        last-lrm-refresh="1281367116"
> ===============
>
> output of monitoring:
> ===============
> Last updated: Mon Aug  9 18:11:20 2010
> Stack: openais
> Current DC: oslo - partition with quorum
> Version: 1.0.9-74392a28b7f31d7ddc86689598bd23114f58978b
> 2 Nodes configured, 2 expected votes
> 5 Resources configured.
> ============
>
> Online: [ oslo tromso ]
>
> Cluster-IP      (ocf::heartbeat:IPaddr2):       Started oslo
>  Clone Set: iscsi-clone
>     Started: [ oslo tromso ]
>  Clone Set: dlm-clone
>     Started: [ oslo tromso ]
>  Clone Set: gfs-clone
>     Started: [ oslo tromso ]
>
> Failed actions:
>    gfs2FS:0_start_0 (node=oslo, call=14, rc=1, status=complete): unknown
> error
>    gfs2FS:0_start_0 (node=tromso, call=12, rc=1, status=complete): unknown
> error
> ===============
>
> and versions of components (all sw manually installed)
>
> tromso:/etc/corosync # dlm_controld.pcmk -V
> dlm_controld 3.0.14 (built Aug  4 2010 23:13:16)
> Copyright (C) Red Hat, Inc.  2004-2010  All rights reserved.
> ===============
> tromso:/etc/corosync # gfs_controld.pcmk -V
> gfs_controld 3.0.14 (built Aug  4 2010 23:13:35)
> Copyright (C) Red Hat, Inc.  2004-2010  All rights reserved.
> ===============
> tromso:/etc/corosync # corosync -v
> Corosync Cluster Engine, version '1.2.7' SVN revision '3013'
> Copyright (c) 2006-2009 Red Hat, Inc.
> ===============
>
> this is kernel...
> Linux version 2.6.34-12-xen (geeko at buildhost) (gcc version 4.5.0 20100604
> [gcc-4_5-branch revision 160292] (SUSE Linux) ).
>
> i'm not a guru and this installation is only for learning so be patient, and
> sorry for my poor english.
> thank to all in advance and nice regards.
>
> PS: i hope this is the right place to post my question, if not please let me
> know.
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs:
> http://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker
>

More information about the Pacemaker mailing list