https://mariadb.com/kb/en/library/server-system-variables/#lock_wait_timeout

Since metadata operations may be long (an ALTER could take a very long time, even days) a high wait time is not a problem, but actually a good thing.  Unless you are seeing lock timeouts that you were not seeing before, you will not see any impact from reducing the lock wait timeout.

Note, that row lock times for row locking engines are engine specific , ie: innodb_row_lock_timeout.

On Fri, Oct 18, 2019, 4:28 AM <jkmdb@w3.org> wrote:
Hi,

When migrating from mariadb 10.1 to 10.3 we noticed that
the default value for lock_wait_timeout went from 1 year
to 1 day.

Although 1 day seems more reasonable than 1 year, it seems
pretty high for a lock_wait.

Does anyone know why this value is so high? Are there any
rule of thumbs for a good timeout value that won't degrade
the system and yet allow it to operate as intended?

I know that part of the answer is "it depends on which
operations your applications are doing". What I don't know
is if beyond those applications, mariadb needs a specific
timeout to function properly, for example, when doing a
mariabackup or flushing tables.

Is anyone else using a timeout value other than the default ones?

My apologies for the naivete of this question.

Thanks!

--josé

_______________________________________________
Mailing list: https://launchpad.net/~maria-discuss
Post to     : maria-discuss@lists.launchpad.net
Unsubscribe : https://launchpad.net/~maria-discuss
More help   : https://help.launchpad.net/ListHelp