The Search indexing pyFTSIncrementalIndexer queue processor have pending search to update, which will take approximately xx hrs
Hi All,
We have recently upgrade our env to 8.5 and post that when items were reindexed in search landing page we have observed that queue size for FTSIncrementalIndexer QP is keep increasing.
Also, receiving PDC alert OPS0014 for the same .
Does any one have any idea on how to resolve this issue.
Pegasystems Inc.
US
@wanig What has likely happened is that one or more of your nodes failed to start elasticsearch, or the indexes are unhealthy for some reason.
Make sure the pyFTSIncrementalIndexer QP is running. If it's not, turn it on and monitor it. If it turns off by itself, it means it's detecting an issue of some sort.
You can check the pyIndexerState column of the pr_sys_statusnodes table to ensure all the nodes have it set to Active. If some are "Initializing" or "Failed", then try restarting those nodes.
Ensure all nodes have the -Dindex.directory JVM argument set. Nodes hosting indexes should have a value specified, nodes that will never host indexes still need the JVM argument, but with no value given. This forces nodes to act either as an index host node or not, which is necessary to avoid issues. If this was not done already, bring down all your nodes, clear out the index directories' contents, set the JVM args, then bring the nodes back up. This may be necessary if the indexes are already in a bad state because the value wasn't set before.