are there any known problems with running multiple transactors and cassandra? i'm seeing my two transactors flip-flop even when i set heartbeat-interval-msec relatively high (at the moment it's at 15000)

i'm also not sure how much of a problem this is, but it seems to be causing connection problems for the peers in our app anyway

@gworley3: on latest version? It wouldn’t be caused by heartbeat latency, but there was a bug fixed in 0.9.5327 that could cause transactor terminations. From changes.md: * Bugfix: Fixed bug that could cause a transactor to terminate when deleting a database against Cassandra storage.

There’s nothing otherwise intrinsic to a Cassandra configuration that would result in that high frequency of failovers. First sanity questions for seeing failovers like that is : * Are processes (storage, transactor, peers) isolated, is network free of partition events * is Cassandra system seeing other use, up to and including compact events, etc. that can cause storage unavailability, Do you see, e.g., the cause of transactor failover listed in logs? I.e., :cause :conflict with AlarmHeartbeatFailed?

@bkamphaus: 0.9.5344. this is on a cluster just for datomic and it has no traffic to it other than the transactors right now. i'm not seeing obvious signs of network partitions. i am seeing :event :transactor/heartbeat-failed, :cause :conflict in the logs

i don't see AlarmHeartbeatFailed anywhere, though

@gworley3: @bkamphaus check JVM GC logs for long pause times as well, both on Cassie and transactor (Cassie only matters if all the nodes sync a long GC together, but never say never)

@gworley3: another sanity check, if they’re flip flopping immediately, does either one ever successfully become a standby (i.e. issue HeartMonitorMsec metric) and related question, is this with a paid or eval license? (no HA is one of the pro starter limitations along with non memcached and 3 process cap).

@bkamphaus: ohhhhhh, that's the answer. this is with a pro starter license right now since we're still in development and not ready to deploy yet. if that's the case i'll check into getting the full thing since we're already past the committed point

i didn't realize it didn't work with ha. Thanks!

Hey, newbie question: is it possible to execute transactions in the datomic console? Would like to set up some test data while I try to comprehend datalog

...or do people generally use the REPL for this sort of thing?

REPL is king

@timgilbert: The Console doesn’t support transactions.

Ok, got it, thanks

The ReST API can do transactions, and it includes basic HTML forms when visited from a browser.

can I treat backward-reference as an attribute?

:community/_neighborhood

@isaac: if you're asking what I think you're asking, yes

@isaac: it depends on where you’re using it. Works in pull/entity, not in query/datalog (but just reorder the from e.g. [?n :community/_neighorhood ?c] to [?c :community/neighborhood ?n]