Tim Coleman

Hi @dpasek-senacor, thanks for the detailed information, we were able to reproduce the problem and have a fix coming through the pipeline https://github.com/EventStore/EventStore/pull/3584

Hi @dpasek-senacor, yes you're right, this is a behaviour we'd like to improve in the near future. You're also right that the 10s is the result of an internal timeout....

Thanks! I've opened a ticket for the slow repsonse here https://github.com/EventStore/EventStore/issues/3617

@mmuller99 in addition to what @hayley-jean has suggested, you could try setting the stream info cache size. the default of `0` causes it to be dynamically sized on startup according...

i pushed a bunch of trivial adjustments to a branch `dynamic-cache-sizeing-tim-suggestions`, consider these suggestions you can merge/squash in whichever ones you like

could update LogV2Sizer to have a more accurate calculation like the memsizer, since we're using the LogV2Sizer to do part of the size calculation

Hi @rastrup @StevenBlair123, i think quite likely https://github.com/EventStore/EventStore/pull/3532 will fix the behaviour you've seen here

Nah i think the new dynamic sizing is just part of it - preventing it from running out of memory. But it may still be far larger by default than...

for `when_requesting_from_follower` i think its this guy `AssertEx.IsOrBecomesTrue(()=> replica.Db.Config.IndexCheckpoint.Read() == leaderIndex);` suggest change to `AssertEx.IsOrBecomesTrue(()=> replica.Db.Config.IndexCheckpoint.Read() >= leaderIndex);`

This should be fixed in the upcoming 22.10.0 release. I'll close this for now please feel free to re-open if you have further problems Thanks!