Hello,
I have a question about updating a solr cloud cluster servers in place. I
have a scripted method for updating a solr cloud in place, which works
consistently to up/down grade between 6.0.0 and 6.0.1 (in our test
environment), but hits an error consistently when going from either to solr
6.1.0. Each server is hosting a single solr node, and each shard has a
replication factor of 3.
The way the script works is as follows. For each instance:
1. Pull the instance from serving requests and drain.
2. Delete the replica from the collection (but leave the index and data)
3. If that node was the leader, force a leader election (solr is not
accepting writes at this time, so this is safe)
4. Run a bootstrapping script on the remote machine (which installs a
particular solr version, and is otherwise idempotent)
5. Once the instance is updated and solr is confirmed, add the node as a
replica where it used to be
6. Wait for recovery
7. Reserve requests from this node.
As mentioned, this hasn't shown any problem switching between versions
6.0.0 and 6.0.1, but when I try to use this to upgrade to solr 6.1.0, the
"ADDREPLICA" command fails as follows:
"status":500,"QTime":65},"failure":{"172.18.6.68:8983_solr":"org.apache.solr.client.solrj.impl.HttpSolrClient$RemoteSolrException:Error
from server at http://172.18.6.68:8983/solr: Expected mime type
application/octet-stream but got text/html
I've included the full log below as an attachment. The exact request being
served is the following:
http://52.91.138.30:8983/solr/admin/collections?wt=json&action=ADDREPLICA&collection=panopto&shard=shard2&node=172.18.6.68:8983_solr
I didn't see any special actions which needed to be taken when upgrading to
6.1.0. Is there perhaps something wrong in my upgrade methodology or
anything else you're aware of which may be related?
Thanks for your help!
Stephen
--
www.stephen-lewis.net
