next local WC can we look at upgrading SOLR from 6.x to 8.x? Under heavy load, the SOLR nodes tend to crash. solr-2 died about 65 minutes ago and has been resyncing with the cluster leader. It's stable under normal load, but reindexes are pushing SOLR 6.x past what it can reliably handle
not to mention when it happens the leader grabs tons of RAM, solr-0 has 31GB of RAM active atm :exploding_head:
Solr 6 is causing cluster instability and resource hogging. @rordway believes an upgrade to solr8 should help smooth out heavy solr loads
Expected behavior
Solr 8.x in development and cluster
Actual behavior
Solr 6.x in development and cluster
Steps to reproduce the behavior
Start reindex-by-chunk job in B8S
See Solr nodes fall over
See Solr leader grab tons of RAM
Descriptive summary
Solr 6 is causing cluster instability and resource hogging. @rordway believes an upgrade to solr8 should help smooth out heavy solr loads
Expected behavior
Solr 8.x in development and cluster
Actual behavior
Solr 6.x in development and cluster
Steps to reproduce the behavior
Start reindex-by-chunk job in B8S See Solr nodes fall over See Solr leader grab tons of RAM
Related work