Hi, We struggle a bit with our cassandra cluster while using the akka-persistence-cassandra journal, in a cluster sharding context with a few thousand short-lived (from 15 min to 15 days) persistent actors, and a few long-lived persistent actors, and 2 persistence views (with a 2s update cycle).
Most if not all the long-lived actors store a snapshot every 15 minutes, and *delete* old events prior to this snapshot at the same time. At start everything works correctly but after some time, the persistent views read requests starts to take more and more time (after something like 1 or 2 days), until they timeout and our cassandra cluster starts going crazy consuming 100% CPU on all nodes (presumably because reads takes so long that they are queued). Our data set is around 80MB, which is very small (running on powerful hardware with SSDs and lots of RAM), which should be essentially fully in RAM. I'm thinking this is related to static columns and tombstones, or the way compaction is run. So I was wondering if anyone had live production tips to share (even answers like "it works like a charm with X or Y workload") would help comfort me. Many thanks for any answers! -- Brice Figureau My Blog: http://www.masterzen.fr/ -- >>>>>>>>>> Read the docs: http://akka.io/docs/ >>>>>>>>>> Check the FAQ: >>>>>>>>>> http://doc.akka.io/docs/akka/current/additional/faq.html >>>>>>>>>> Search the archives: https://groups.google.com/group/akka-user --- You received this message because you are subscribed to the Google Groups "Akka User List" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/akka-user. For more options, visit https://groups.google.com/d/optout.
