These are chat archives for YoEight/eventstore

12th
Jun 2017
Akii
@Akii
Jun 12 2017 11:37
hm getting some heartbeat related disconnects attached with the driver sending events multiple times
can I fix that by adjusting s_heartbeatInterval and/or s_heartbeatTimeout?
Yorick Laupa
@YoEight
Jun 12 2017 12:05

Those aren't used right now. The driver solely depends on server side heartbeat protocol (might add it for the next version). So those would have to be set on the server itself for now.

Before that. Are you under a heavy workload ? If yes, it might be a good call to increase those values.

Also, make sure you'll also running your application with -threaded GHC options. If under heavy workload, the driver will be slower if it doesn't run on all your cores

Akii
@Akii
Jun 12 2017 12:12
nope, the thing is idling
might be a very slow server
I really don't get these "slow bus msg" log entries
ghc-options: -threaded -rtsopts -with-rtsopts=-N
should be good
Akii
@Akii
Jun 12 2017 12:17
OT: do you remember my projection that was so slow (took around 5-10 minutes; doing a lot of nested record updates)? Rewrote that, slightly changed the structure and used lens for updating. Now runs in 1 second.
(~130k events)
but re this "issue": should I investigate this or just wait for the major refactoring; it's probably already fixed there
I hate this uncertainty about whether I fucked up the server, the app, the store or if it's something in the driver xD
Yorick Laupa
@YoEight
Jun 12 2017 13:18
the Slow bus msg is when the server is processing message from its multiple concurrent queue
generally it's when it takes more than 200ms
but it changes depending on which queue is it
that's not an issue
Akii
@Akii
Jun 12 2017 14:58
well might be related to the getEventStore update I did 3.9.3 -> 3.9.4 afaik
Akii
@Akii
Jun 12 2017 16:26
so I'll just implement a naive form of deduplication
guess that works
kinda ugly
Akii
@Akii
Jun 12 2017 20:30
btw, you on Slack or Discord?