All,
- I'm have a small server I'm testing with some concepts for SQL
integration with MD
- MD is set to a maximum of 45 slaves.
- 98.143% of the processes are handled with the top 10 slaves. 99.591%
of the processes handled with the top 15 slaves.
- I have a mysql server with a 3600 second (1 hour) timeout in the my.cnf
- I have the mysql client connecting to the mysql server
- I have a tick command in my filter
- Multiplexor is running with -X 60 -P 3
- My tick command ignores the banding so every tick type runs the same
function which essentially is a dbh->ping and an SQL reconnection if the
ping fails.
- In the tick command, I see pings fail to the DB server occasionally .
Yesterday , 4281 successful pings, 14 unsuccessful.
- This race condition sometimes leads to queries being run where the
connection has been lost to the database.
My theory is that I should be able to tweak -P and -X to run more ticks
in parallel to effectively end the client timeout unless the slave is
gracefully killed.
I'm also rewriting my db query function so that if the query fails, it
will try and run it again a 2nd time after reconnecting. However, that's
a band-aid IMO.
Anyone have any guidance on settings for -P and -X in order to maintain
persistent DB connections?
Regards,
KAM
_______________________________________________
NOTE: If there is a disclaimer or other legal boilerplate in the above
message, it is NULL AND VOID. You may ignore it.
Visit http://www.mimedefang.org and http://www.roaringpenguin.com
MIMEDefang mailing list [email protected]
http://lists.roaringpenguin.com/mailman/listinfo/mimedefang
