[Pacemaker] Antw: Re: gfs2: ping_pong test fails

[Andrew Beekhof]

On Fri, Feb 26, 2010 at 4:11 PM, Raphael Daum <Raphael.Daum at kurier.at> wrote:
> dear all,
>
> thanks for valuable feedback.
>
>>>> Ante Karamatic <ivoks at ubuntu.com> schrieb am 24.02.2010 um 09:25:
>> On 22.02.2010 21:23, Andrew Beekhof wrote:
>>
>>> Wait a second, thats not right. You want:
>>>
>>> primitive gfsd ocf:pacemaker:controld
>>>       params daemon=gfs_controld.pcmk args="-g 0"
>>>       op monitor interval=120s
>
> thanks for the pointer. i updated the ra to support it.

the RA or the resource definition?
the RA should not have needed modification

> i tried to
> reproduce the simple setup to test the functionality, skipping the clvm
> resource and running the ping_pong test against a "normal" partition.
> here my config.
>
> node node1
> node node2
> primitive data-test ocf:heartbeat:Filesystem \
>        params device="/dev/mapper/mpathcp1" directory="/data/test"
> fstype="gfs2" \
>        op monitor interval="120s"
> primitive dlm ocf:pacemaker:controld \
>        op monitor interval="120s" \
>        meta target-role="Started"
> primitive gfsd ocf:pacemaker:controld \
>        params daemon="gfs_controld.pcmk" args="-g 0" \
>        op monitor interval="120s" \
>        meta target-role="Started"
> clone data-test-clone data-test \
>        meta interleave="true" ordered="true"
> clone dlm-clone dlm \
>        meta interleave="true" ordered="true"
> clone gfsd-clone gfsd \
>        meta interleave="true" ordered="true"
> colocation colo-gfsd-WITH-dlm inf: gfsd-clone dlm-clone
> order order-data-test-AFTER-gfsd inf: gfsd-clone data-test-clone
> order order-gfsd-AFTER-dlm inf: dlm-clone gfsd-clone

You'd also want:
  colocation colo-fs-with-gfsd inf: data-test gfsd-clone

but that shouldn't have caused a problem (assuming it was able to start up ok)

>
> as may imagine, the ping_pong test still fails. stracing the ping_pong
> test gives the followinig output:
>
> write(1, "lock at 0 failed! - Function not"..., 45) = 45
> fcntl(3, F_SETLKW, {type=F_UNLCK, whence=SEEK_SET, start=0, len=1}) =
> -1 ENOSYS (Function not implemented)
> write(1, "unlock at 0 failed! - Function n"..., 47) = 47
> fcntl(3, F_SETLKW, {type=F_WRLCK, whence=SEEK_SET, start=0, len=1}) =
> -1 ENOSYS (Function not implemented)
>
> running the ping_pong test on a local filesystem (ext3 in this case)
> works. if i mount manually mount the gfs2-partition with the mount
> option "localflocks" again the test succeds.
>
> since the test works for other people, i'm rather clueless what i'm
> doing wrong. here some details about my system:
>
> # uname -a
> Linux node1 2.6.31.12-174.2.19.fc12.x86_64 #1 SMP Thu Feb 11 07:07:16
> UTC 2010 x86_64 x86_64 x86_64 GNU/Linux
>
> # packages (all taken from the fedora repos)
> cluster-glue-libs-1.0-0.11.b79635605337.hg.fc12.x86_64
> clusterlib-3.0.8-1.fc12.x86_64
> cluster-glue-1.0-0.11.b79635605337.hg.fc12.x86_64
> corosynclib-1.2.0-1.fc12.x86_64
> corosync-1.2.0-1.fc12.x86_64
> pacemaker-1.0.7-1.fc12.x86_64
> pacemaker-libs-1.0.7-1.fc12.x86_64
> resource-agents-3.0.7-2.fc12.x86_64
> dlm-pcmk-3.0.8-1.fc12.x86_64
> gfs-pcmk-3.0.8-1.fc12.x86_64
> gfs2-utils-3.0.8-1.fc12.x86_64

all those versions look reasonable, i'm really not sure whats going on.
perhaps dlm_controld has more information in its debug logs (though I
forget how to tell it to dump that out) - the linux-cluster list
should know.

>
> # dlm_controld.pcmk -V
> dlm_controld 3.0.8 (built Feb 23 2010 13:00:42)
>
> #  gfs_controld.pcmk -V
> gfs_controld 3.0.8 (built Feb 23 2010 13:00:50)
>
> any advices? maybe this a problem with this specific kernel version?
>
> TIA
>
> raphael
>
> Telekurier Online Medien GmbH & Co KG
> 1070 Wien, Lindengasse 52
> FN 216605m, HG Wien
> UID: ATU53024506
> Komplementär: Telekurier Online Medien GmbH, FN209236a, HG Wien
>
> ------------------------------------------------------------------------------------------------
>
> Diese E-Mail enthaelt vertrauliche und/oder rechtlich geschuetzte
> Informationen.
> Wenn Sie nicht der richtige Adressat sind oder diese E-Mail irrtuemlich
> erhalten
> haben, informieren Sie bitte sofort den Absender und vernichten Sie
> diese Mail.
> Das unerlaubte Kopieren sowie die unbefugte Weitergabe dieser Mail ist
> nicht
> gestattet.
>
> The information transmitted is intended only for the person or entity to
>
> which it is addressed and may contain confidential and/or privileged
> material. Any review, retransmission, dissemination or other use of, or
> taking of any action in reliance upon, this information by persons or
> entities
> other than the intended recipient is prohibited. If
> you received this in
> error,
> please contact the sender and delete the material from any computer.
> ------------------------------------------------------------------------------------------------
>
>
> _______________________________________________
> Pacemaker mailing list
> Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>