Re: Fwd: Postgres HA - pacemaker RA do not support auto failback

On Tue, 13 Aug 2019, 11:50 Shital A, <brightuser2019(at)gmail(dot)com> wrote:

>
>
> On Mon, 12 Aug 2019, 18:54 Adrien Nayrat, <adrien(dot)nayrat(at)anayrat(dot)info>
> wrote:
>
>> On 8/12/19 2:57 PM, Shital A wrote:
>> >
>> >
>> > Hello,
>> >
>> > Postgres version : 9.6
>> > OS:Rhel 7.6
>> >
>> > We are working on HA setup for postgres cluster of two nodes in
>> active-passive
>> > mode.
>> >
>> > Installed:
>> > Pacemaker 1.1.19
>> > Corosync 2.4.3
>> >
>> > The pacemaker agent with this installation doesn't support automatic
>> failback.
>> > What I mean by that is explained below:
>> > 1. Cluster is setup like A - B with A as master.
>> > 2. Kill services on A, node B will come up as master.
>> > 3. node A is ready to join the cluster, we have to delete the lock file
>> it
>> > creates on any one of the node and execute the cleanup command to get
>> the node
>> > back as standby
>> >
>> > Step 3 is manual so HA is not achieved in real sense.
>> >
>> > Please help to check:
>> > 1. Is there any version of the resouce agent which supports automatic
>> failback?
>> > To avoid generation of lock file and deleting it.
>> >
>> > 2. If there is no such support, what checks should be added in pgsql RA
>> to achieve
>> >
>> > Please suggest.
>> > Thanks.
>> >
>>
>> Hello,
>>
>> Which RA did you use? AFAIK there is two RA :
>> - pgsql :
>> https://github.com/ClusterLabs/resource-agents/blob/master/heartbeat/pgsql
>> - PAF : http://clusterlabs.github.io/PAF/ (Which is different from pgsql
>> RA :
>> http://clusterlabs.github.io/PAF/FAQ.html#why-new-ra-for-postgresql)
>>
>> If I am right, PAF should handle the case when there no need to perform
>> pgrewind
>> (if the old master is not too advanced in the transaction log).
>>
>> Regards,
>>
>> --
>> Adrien NAYRAT
>> https://blog.anayrat.info
>
>
>
>
> Hello
>
> We are using the pgsql RA installed with pacemaker.
>
> Is PAF recommended over pgsql?
>
> Has anyone changed pgsql to handle the lock file and other cases that
> might not have been handled?
>
> Please advise.
>
> Thanks.
>

Looks like the previous email bounced, received undelivered message.
Sending again. We are using the pgsql RA installed with pacemaker and seems
like it prevents failback by creating a lock file. We want to implement
auto failback and need pointers on how to implement.

Thanks.