[Archivesspace_Users_Group] Implementing a standalone Solr instance

[Peter Heiner]

Hi Kyle,

We at the Cambridge University Library have been running an external Solr
instance since the early days of our deployment.
We use Puppet for configuration management of our local VM stack and the
module we use is https://forge.puppetlabs.com/modules/landcareresearch/solr.

We currently store over 13K resources comprised of over 900K archival objects
across 30 repositories, the number of attached notes is well above 2.2M.
It's important to note that our initial archival object count was over 700K, I
believe we were fairly large scale users even initially.

Our production Solr currently has just under 3 million documents in the index.
Availability and performance of Solr is, in our experience, the single most
common cause of ArchivesSpace downtime, ahead of Java memory leakage, planned
maintenance due to upgrades, and database or network failures.

For our scale and use, especially while we were in the migration phase
occasionally bringing in hundreds of thousands of archival objects, we have
found that indexing threads could also interfere with interactive use of the
site and after some unsuccessful attempts at fixing this by altering the
indexer thread count we have split the indexer off to another VM. Our Ops team
have recently identified and fixed issues with our VM infrastructure, so we're
waiting for more monitoring data to confirm whether we still need this, but I
imagine having Solr, AS, and the indexer on the same VM might hurt
performance.

Hope that helps,
p

Kyle Breneman wrote on 2021-12-03 16:50:29:
> Here at the University of Baltimore, we are working to stand up our own external Solr installation in preparation for ArchivesSpace's move away from bundled Solr.  Campus IT is asking me whether or not we should install Solr on the same server we're using to host ArchviesSpace.  I see that ArchivesSpace officially has no opinion<https://archivesspace.org/archives/7137> on this matter (See under "Will you have strict requirements for how to deploy Solr?").
> 
> Does anyone on this list have a recommendation about whether an external instance of Solr should be installed on the same server as ArchivesSpace, or on its own separate server?  I don't see that it makes much difference, but I am also not experienced in managing servers, or in administering Solr.
> 
> Kyle Breneman
> Integrated Digital Services Librarian
> The University of Baltimore
> kbreneman at ubalt.edu<mailto:kbreneman at ubalt.edu>
> I believe in freedom of thought and
> freedom of speech. Do you?
> 

> _______________________________________________
> Archivesspace_Users_Group mailing list
> Archivesspace_Users_Group at lyralists.lyrasis.org
> http://lyralists.lyrasis.org/mailman/listinfo/archivesspace_users_group