[Openais] Help on mounting ocfs2 filesystems

Andreas Kurz andreas at hastexo.com
Wed Mar 21 22:40:51 UTC 2012


On 03/21/2012 09:50 PM, Andreas Kurz wrote:
> On 03/21/2012 08:46 PM, Carlos Xavier wrote:
>> Hi.
>>
>> You just gave  me the silver bullet :-)
>>
>>> Hello,
>>>
>>> ... see comments inline ...
>>>
>>> On 03/21/2012 03:24 PM, Carlos Xavier wrote:
>>>> Tank you for the quick anser
>>>>>>
>>>>>> Im tryng to make a cluster using pacemaker on OpenSUSE 12.1 with
>>>>>> DRBD +
>>>>>> OCFS2 + Mysql on top of the filesystem.
>>>>>> The system will have two DRBD resources to be mounted on
>>>>>> /var/lib/mysql
>>>>>> and on /export.
>>>>>
>>
>> snip
>>
>>>> colocation colDLMDRBD inf: cloneDLM msDRBD_export:Master
>>>> colocation colFSO2CB inf: cloneFSexport cloneO2CB
>>>> colocation colO2CBDLM inf: cloneO2CB cloneDLM
>>>> order ordDLMO2CB 0: cloneDLM cloneO2CB
>>>> order ordDRBDDLM 0: msDRBD_export:promote cloneDLM
>>>
>>> .. should be cloneDLM:start
>>>
>>
>> Fixed
>>
>>>> order ordO2CBFS 0: cloneO2CB cloneFSexport
>>>> property $id="cib-bootstrap-options" \
>>>>        dc-version="1.1.6-9971ebba4494012a93c03b40a2c58ec0eb60f50c" \
>>
>> snip
>>
>>>> (resFSexport:0:start:stderr) mount.ocfs2: Cluster stack specified does
>>>> not match the one currently running while trying to join the group
>>>
>>> You created the ocfs2 file system without Pacemaker running? You need to
>>> do a: tunefs.ocfs2 --update-cluster-stack <device>
>>>
>>
>> This was the case, as it is my first time with Pacemaker, I didnt have
>> any clues that the
>> communication between the servers is replaced with the Pacemaker stack.
>> I did what you told me and now both filesystems are mounted.
>>
>> Now I have some doubts. The configuration:
>>
>> order ordDRBDDLM 0: msDRBD_export:promote cloneDLM:start
>>
>> makes sence with just one mount point, since the start of the resource
>> cloneDLM:start is tied to the msDRBD_export:promote. If msDRBD_export
>> fails, msDRBD_mysql cant be mounted, since the DLM and OCFS2 wont be
>> available.
> 
> Yes, for several DRBD resources I'd recommend you let each file system
> mount depend on its DRBD resource (order/colocation), as well as on
> dlm/oc2b ... as dlm/o2cb have no dependency on DRBD but the ocfs2 mounts
> depend on locking and DRBD.
> 
>>
>> How can I change this statement to get the cloneDLM started having
>> msDRBD_export or msDRBD_mysql started?
>> Can I use one resource more than one time? Because i dont have any
>> colocation or order statement for mounting /var/lib/mysql.

Just to mention it: it makes no sense to put mysql data files on a
clustered files system. You can't run two mysql prozesses concurrently
on the same data set ....

There are really very few and specialized use cases to run a cluster
file system like OCFS2 at all. Most scenarios are really fine with a
nice and lightweight active-passive high availability cluster setup.

Regards,
Andreas

-- 
Need help with Pacemaker?
http://www.hastexo.com/now

> 
> of course, you can reference a resource in several constraints
> 
>>
>> primitive ip_mysql ocf:heartbeat:IPaddr2 \
>>        params ip="10.10.10.5" cidr_netmask="32" nic="vlan0" \
>>        op monitor interval="30s"
>> primitive resDLM ocf:pacemaker:controld \
>>        op monitor interval="60" timeout="60"
>> primitive resDRBD_export ocf:linbit:drbd \
>>        params drbd_resource="export" \
>>        operations $id="opsDRBD_export" \
>>        op monitor interval="20" role="Master" timeout="20" \
>>        op monitor interval="30" role="Slave" timeout="20" \
>>        meta target-role="started"
>> primitive resDRBD_mysql ocf:linbit:drbd \
>>        params drbd_resource="mysql" \
>>        operations $id="opsDRBD_mysql" \
>>        op monitor interval="20" role="Master" timeout="20" \
>>        op monitor interval="30" role="Slave" timeout="20" \
>>        meta target-role="started"
>> primitive resFSexport ocf:heartbeat:Filesystem \
>>        params device="/dev/drbd/by-res/export" directory="/export"
>> fstype="ocfs2" options="rw,noatime" \
>>        op monitor interval="120s"
>> primitive resFSmysql ocf:heartbeat:Filesystem \
>>        params device="/dev/drbd/by-res/mysql" directory="/var/lib/mysql"
>> fstype="ocfs2" options="rw,noatime" \
>>        op monitor interval="120s"
>> primitive resO2CB ocf:ocfs2:o2cb \
>>        op monitor interval="60" timeout="60"
>> ms msDRBD_export resDRBD_export \
>>        meta resource-stickines="100" master-max="2" clone-max="2"
>> notify="true" interleave="true"
>> ms msDRBD_mysql resDRBD_mysql \
>>        meta resource-stickines="100" master-max="2" clone-max="2"
>> notify="true" interleave="true"
>> clone cloneDLM resDLM \
>>        meta globally-unique="false" interleave="true"
>> clone cloneFSexport resFSexport \
>>        meta interleave="true" ordered="true"
>> clone cloneFSmysql resFSmysql \
>>        meta interleave="true" ordered="true"
>> clone cloneO2CB resO2CB \
>>        meta globally-unique="false" interleave="true"
>> colocation colDLMDRBD inf: cloneDLM msDRBD_export:Master
>> colocation colFSO2CB inf: cloneFSexport cloneO2CB
>> colocation colO2CBDLM inf: cloneO2CB cloneDLM
>> order ordDLMO2CB 0: cloneDLM cloneO2CB
>> order ordDRBDDLM 0: msDRBD_export:promote cloneDLM:start
>> order ordO2CBFS 0: cloneO2CB cloneFSexport
>>
>>>
>>> Regards,
>>> Andreas
>>>
>>
>> Thank you very much for your help and cooperation.
> 
> You are welcome!
> 
> Regards,
> Andreas
> 
> 
> 
> 
> _______________________________________________
> Openais mailing list
> Openais at lists.linux-foundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/openais




-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 222 bytes
Desc: OpenPGP digital signature
URL: <http://lists.linuxfoundation.org/pipermail/openais/attachments/20120321/e4580a41/attachment.sig>


More information about the Openais mailing list