[
https://issues.apache.org/jira/browse/SOLR-8129?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15111833#comment-15111833
]
Mark Miller commented on SOLR-8129:
-----------------------------------
bq. we're saying the same thing.
I just get the feeling you have been trying to avoid doing this full hash, but
it seems so much more bulletproof than these quick window checks. I'd rather
recovery be a lot slower and get it right and worry about speeding that up
later.
bq. the only current general remedy we have is to fail peersync and copy the
whole index.
These kind of reorders don't seem that common? And we have many other reasons
you will probably have to replicate the whole index. 100 updates in peer sync
means it's best simply for static indexes or low usage indexes. Sure, I'd
rather not replicate a full index, but I think it's pretty common to have to do
already.
> HdfsChaosMonkeyNothingIsSafeTest failures
> -----------------------------------------
>
> Key: SOLR-8129
> URL: https://issues.apache.org/jira/browse/SOLR-8129
> Project: Solr
> Issue Type: Bug
> Reporter: Yonik Seeley
> Attachments: fail.151005_064958, fail.151005_080319
>
>
> New HDFS chaos test in SOLR-8123 hits a number of types of failures,
> including shard inconsistency.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
