Hi Experts,
We have TarMK Cold standby setup for our Author environment (AEM 6.3 SP3) and facing Consistency issue on standby server, please suggest if someone have similar experience -
Background:
- The standby is created by copying the installation directory of Primary.
- We have external NAS storage for segmentStore and dataStore so, I copied that too while creating the standby environment.
- Before starting the standby, deleted the sling.id.file from installation directory and assured that dataStore directory is having only one repository_<UUID> file which is exactly same as Primary file (as it was copied from Primary)
- Followed all instruction given in adobe documentation and additionally turned on the autocleanup for standby instance.
- Everything is working great – content/assets sync is happening properly, validated through logs and by restarting the standby in primary mode.
Issue:
- After setting up my standby if I run the jmx-->BlobGarbageCollection-->checkConsistency(), it shows 0 missing blobs which is expected. But, as soon as sync start between Primary and Standby even after hours/days whenever I run the checkConsistency() on standby server it always shows me 4K, 5k etc. missing blobs. The number of missing blobs is varying some time it comes down nominally but mostly it is increasing.
- There is no error in error logs (like segment missing or record doesn’t exist etc.)
- JMX console config of Primary and Standby is clearly showing there is no failed request and segments and binaries are being transferred successfully. Same I can see in tarmk-coldstandby.log as well.
- Error.logs are showing below AuthenticationSupport service missing error but based on earlier ticket with Day Care, it was communicated that because there is no segment missing or record doesn’t exist kind of errors and sync is happening so we can ignore it.
22.11.2018 09:11:19.099 *ERROR* [qtp298931566-1785] org.apache.sling.engine.impl.SlingHttpContext handleSecurity: AuthenticationSupport service missing. Cannot authenticate request.
22.11.2018 09:11:19.100 *ERROR* [qtp298931566-1785] org.apache.sling.engine.impl.SlingHttpContext handleSecurity: Possible reason is missing Repository service. Check AuthenticationSupport dependencies.
However, if I start my standby in Primary mode, I am not getting AuthenticationSupport service missing error which is good sign.
Please suggest if both items (consistency issue, AuthenticationSupport service missing) are expected on Standby server, couldn’t find it in any documentation.
Thanks & Regards,
Deepak Jangra
