I'm trying to upgrade my Solr cluster from 5.4 to 5.5.2 in a rolling restart without success. The first node upgraded with 5.4 worked without any issue, the problem arise with the second. When the second node is restarted with the 5.4 version, the heap of both nodes grows until the first node (I don't why, but always is the first node) hit an OOM. It's like something in the PeerSync process is consuming ram (my index is huge I have replicas with 250G) until hit the OOM.
https://issues.apache.org/jira/browse/SOLR-8586 was added in 5.5, this issue changes the way how shard synchronization is done. Can this issue be related with my problem? \-- /Yago Riveiro ![](https://link.nylas.com/open/m7fkqw0yim04itb62itnp7r9/local-2eefb82c- 7bf1?r=c29sci11c2VyQGx1Y2VuZS5hcGFjaGUub3Jn)