David Smiley created SOLR-8791:
----------------------------------
Summary: ManagedIndexSchema has race of schemaZkVersion causing
BadVersionException
Key: SOLR-8791
URL: https://issues.apache.org/jira/browse/SOLR-8791
Project: Solr
Issue Type: Bug
Environment: 7.0 snapshot
Reporter: David Smiley
I started SolrCloud using a single node, 2 shard, 2 replica per shard setup
with the data driven schema config. I used {{bin/solr start -c -e cloud}} to
do this, although I got the same ultimate problem when doing the setup more
manually. Then I used curl to POST a large CSV file of data. At the end, I
looked in my logs and I saw this:
{noformat}
465889 INFO (qtp680576081-45) [c:geonames s:shard1 r:core_node4
x:geonames_shard1_replica2] o.a.s.s.ManagedIndexSchema Persisted managed schema
version 1 at /configs/data_driven_schema_configs/managed-schema
465889 ERROR (qtp680576081-15) [c:geonames s:shard2 r:core_node3
x:geonames_shard2_replica2] o.a.s.s.ManagedIndexSchema Bad version when trying
to persist schema using 0 due to:
org.apache.zookeeper.KeeperException$BadVersionException: KeeperErrorCode =
BadVersion for /configs/data_driven_schema_configs/managed-schema
{noformat}
When viewing my logs in the admin UI, I actually saw the error message around
~5 times or so in a row, some from the very same replica.
disclaimer: I'm using the latest 7.0 snapshot master branch with some patches
that appear innocently unrelated to this
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
