[
https://issues.apache.org/jira/browse/LUCENE-9576?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17214229#comment-17214229
]
Uwe Schindler commented on LUCENE-9576:
---------------------------------------
I can confirm that any type of stale file mount causes issues:
- Elasticsearch won't start as it does the spinning disk check on startup
- the first Lucene commit hangs.
The bad thing is: it can be caused by any mount, also ones not actually used by
indexing at all. And that's very bad.
I recognized this problem already on PANGAEA where we have some Tape-backed
mounts with huge amounts of data. This affects Lucene, which was completed
unrelated (e.g. when tape roboter changing tapes in background, those mounts
may hang).
> IndexWriter::commit() hangs when the server has a stale NFS mount.
> ------------------------------------------------------------------
>
> Key: LUCENE-9576
> URL: https://issues.apache.org/jira/browse/LUCENE-9576
> Project: Lucene - Core
> Issue Type: Bug
> Components: core/index
> Affects Versions: 8.5.2
> Reporter: Girish Nayak
> Priority: Major
> Attachments: IndexWriter-commit.PNG
>
>
> Noticed IndexWriter::commit() hangs when the server has one or more stale NFS
> mounts.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org
