Closed bc-ashokmahajan closed 2 years ago
I am not sure if it is directly related to above issue, we have consistently seen that the memory footprint of the proxysql process gradually increase over the time and it reduces only when proxysql instance is restarted. Following are grafana graphs of the two memory related metrics for the period of last 10 days.
Same graph for last 1 on hour after the restart:
Your application seems to generate a lot of unique queries, this the memory structure behind stats_mysql_query_digest
is really large. Querying that table causes a lot of processing, and ProxySQL is aborting because it was processing data for really a long period of time, and it is configured to abort if threads make no progress for 30 seconds.
Therefore it aborted as configured.
Today one of our production proxysql node crashed while querying stats_mysql_query_digest_reset, which we were executing after a long time. It resulted into errors to the end users.
(Messages related to update_mysql_servers.sh above are from out scheduler job, so please ignore it.)
Proxysql version:
2.3.2-10-g8cd66cf
The package used to install ProxySQL
We installed it from proxysql repo: https://repo.proxysql.com/ProxySQL/proxysql-2.3.x/buster/
Detail of OS:
Linux 4.19.0-17-cloud-amd64 #1 SMP Debian 4.19.194-3 (2021-07-18) x86_64 GNU/Linux
Core dump, proxysql.log and systems.log:
We would like to share it securely. I sent a dropbox link to proxysql support.
H/W details:
Complete hw information also included in the dropbox folder.