SegmentStore clean up in TarMK cold standby cluster

[Kunal_Gaba_]

As per 6.1 release notes there is a known issue with the auto clean up of segment store in cold standby TarMK cluster. And the recommendation is to restart before running the cleanup. Does this mean that we should schedule manual restarts before the auto clean runs ? How can we schedule that as the standby auto clean job is triggered automatically when the size of the store increases 25% over sync cycle ? Or the recommendation is not to use auto clean and run offline compaction ? 

• Oak TarMK Cold Standby: standby.autoclean removes too many segments. Recommendation is to restart the standby prior to running cleanup (CQ-40862)

[smacdonald2008 · Accepted solution]

Take a look at Jayan's blog - it provides more helpful details:

http://cq-ops.tumblr.com/post/102621710969/how-to-configure-an-aem-6-cold-standby-failover

[Kunal_Gaba_]

Thanks Scott. I have gone through these steps and have no issues with setting up the cluster. The only thing which is not clear is clean up of disk space on stand by node. There is a known issue(mentioned in my initial post) with auto clean job on standby but we don't know what is the workaround. By default standby.autoclean flag is turned off and we don't know whether we should turn it on and just restart the instance every day or at specific times. 

[Opkar_Gill]

Hi Kunal,

how did you proceed with this in the end? I was looking at this topic for a client and saw there is documentation for standby maintenance[1]

Regards,

Opkar

[1]https://docs.adobe.com/docs/en/aem/6-0/deploy/recommended-deploys/tarmk-cold-standby.html#Cold Standby Repository Maintenance

[Kunal_Gaba_]

 Yes we raised support ticket with daycare and I think they added these steps on this document now. We are now following the same steps in our project.