FW: Recovering from machine failure

-----Original Message-----
From: Andy C [mailto:andycsolr@gmail.com]
Sent: 17 March 2018 01:54
To: solr-user <solr-user@lucene.apache.org>
Subject: Recovering from machine failure

Running Solr 7.2 in SolrCloud mode with 5 Linux VMs. Each VM was a single
shard, no replication. Single Zookeeper instance running on the same VM as
one of the Solr instances.

IT was making changes, and 2 of the VMs won't reboot (including the VM where
Zookeeper is installed). There was a dedicated drive which Solr (and
Zookeeper for the one node) where installed on, and a dedicated drive where
the Solr indexes were created.

They believe the drives are still good. Their plan is to create 2 new VMs
and attach the drives from the old VMs to them. However the IP addresses of
the new VMs will be different.

In the solr.in.sh I had set the SOLR_HOST entry to the IP address of the VM.
Is this just an arbitrary name? Will Zookeeper still recognize the Solr
instance if the SOLR_HOST entry doesn't match the IP address.

Obviously I will need to adjust the ZK_HOST entries on all nodes to reflect
the new IP address of the VMs. But will that be sufficient?

Appreciate any guidance.

Thanks
- Andy -