Allessandro,
1) in-place upgrade? have you re-indexed or migrated the old indexes?
--A: Upgraded SOLR on the exiting servers utilizing the current indexes. We
have plans to reindex and flatten the structure but we're not ready for
that yet. With the release of SOLR 9 we need to get to a supported version
now. We had planned on reindexing when we go to SOLR 9
2) Adding or modifying content -> means pushing new documents or updating
documents and then committing? or just pushing the docs causes the crash?
-- Adding new content does not seem to cause the crash. Updating existing
indexes appears to be what is causing the issue.
3) The writes work and the changes are available in the other nodes ->
interesting, so the documents are forwarded but the leader dies after
forwarding them? are you using NRT replicas?
-- A: The cluster is using NRT replicas.
4) Solr attempting to reload a cache with too much info? mmm, do you mean
auto warming the cache? what is your cache configuration?
--A: Filter,Document, and Query cache all default. Filter and Query cache
size and initialsize=0. Document cache size and initialsize=512
-- Custom Cache - cache name="perSegFilter"
class="solr.search.LRUCache"
size="10"
initialSize="0"
autowarmCount="10"
regenerator="solr.NoOpRegenerator"
5) how much frequently do you commit? (hard or soft) ? auto-commit?
-- Autocommit
Thank you
On Tue, May 24, 2022 at 6:16 PM Alessandro Benedetti <[email protected]>
wrote:
> Hi Paul,
> 1) in-place upgrade? have you re-indexed or migrated the old indexes?
> 2) Adding or modifying content -> means pushing new documents or updating
> documents and then committing? or just pushing the docs causes the crash?
> 3) The writes work and the changes are available in the other nodes ->
> interesting, so the documents are forwarded but the leader dies after
> forwarding them? are you using NRT replicas?
> 4) Solr attempting to reload a cache with too much info? mmm, do you mean
> auto warming the cache? what is your cache configuration?
> 5) how much frequently do you commit? (hard or soft) ? auto-commit?
>
> Cheers
> --------------------------
> *Alessandro Benedetti*
> CEO @ Sease Ltd.
> *Apache Lucene/Solr Committer*
> *Apache Solr PMC Member*
>
> e-mail: [email protected]
>
>
> *Sease* - Information Retrieval Applied
> Consulting | Training | Open Source
>
> Website: Sease.io <http://sease.io/>
> LinkedIn <https://linkedin.com/company/sease-ltd> | Twitter
> <https://twitter.com/seaseltd> | Youtube
> <https://www.youtube.com/channel/UCDx86ZKLYNpI3gzMercM7BQ> | Github
> <https://github.com/seaseltd>
>
>
> On Tue, 24 May 2022 at 23:02, Paul Russell <[email protected]> wrote:
>
> > We have done an in place upgrade of a SOLR 7.7.2 cluster to SOLR 8.11.1.
> > using openJDK 8.
> >
> > The core loads fine and searches work as expected. Adding or modifying
> > content causes the leader node to crash with OOM error. No
> > additional errors in the SOLR logs. The writes work and the changes are
> > available in the other nodes.
> >
> > This is a large core (>1Tb).
> >
> > Has anyone seen this issue before? Possible issue with SOLR attempting to
> > reload a cache with too much info?
> >
> > Thanks for time,
> >
> > Paul
> >
>
--
Paul
Russell
VP Integration/Support Services
[image: <!--company-->] <https://www.qflow.com/>
*main:* 314.968.9906
*direct:* 314.255.2135
*cell:* 314.258.0864
9317 Manchester Rd.
St. Louis, MO 63119
qflow.com <https://www.qflow.com/>
