This repository was archived by the owner on May 18, 2021. It is now read-only.
This repository was archived by the owner on May 18, 2021. It is now read-only.
Need to limit DB connections at the persistence layer, especially during reload/replay #104
Description
I've a DB with a bunch of active graphs in it - a few thousand.
During replay at a shard restart, I'm seeing the following:
time="2017-11-08T15:57:39Z" level=info msg="Created graph actor supervisor-0/0246f0d7-da44-4de1-97e9-30afd2d6d998" graph_id=0246f0d7-da44-4de1-97e9-30afd2d6d998 logger=supervisor-0
time="2017-11-08T15:57:39Z" level=info msg="Created graph actor supervisor-0/35203cfc-7c0e-4216-847e-7bcc1869fa63" graph_id=35203cfc-7c0e-4216-847e-7bcc1869fa63 logger=supervisor-0
time="2017-11-08T15:57:39Z" level=error msg="Error getting snapshot value from DB Error 1040: Too many connections" actor_name=supervisor-0/7a9f6777-ce90-4dc8-89c3-3846e83b0fac logger=persistence
time="2017-11-08T15:57:39Z" level=error msg="Error getting snapshot value from DB Error 1040: Too many connections" actor_name=supervisor-0/71c3eba2-43f0-41e6-a9a5-1ef836b74513 logger=persistence
time="2017-11-08T15:57:39Z" level=error msg="Error getting snapshot value from DB Error 1040: Too many connections" actor_name=supervisor-0/f4602309-139e-4fc5-b226-6f33a7528346 logger=persistence
Error 1040 indicates too many parallel connections to the DB. It looks like we're generating these in parallel.