No way to increase the net_*_timeout values for the mysqldump-lvm mysqld process
Bug #1278855 reported by
Marc - A. Dahlhaus
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
holland-backup |
New
|
Undecided
|
Unassigned |
Bug Description
There is currently no way to set the net-read-timeout and net-write-timeout values in the mysqld process dumping contents from a snapshot.
We are using a NFS share that is provided by a NAS that does inline deduplication of the data.
We get "connection lost during query" errors while running the mysqldump-lvm plugin.
I could fix this by adding a way to set the timeouts for the snapshot mysqld process.
To post a comment you must log in.
This is similar to the request in https:/ /bugs.launchpad .net/holland- backup/ +bug/752048 where a more general mysqld-options config parameter is being added to allow passing arbitrary options to the mysqld commandline. Probably allow for additional options in the holland backupset's [mysqld] section also makes sense here.
I am surprised that net-*-timeout was causing a problem here - those are pretty high at 30 and 60 seconds by default. Did this resolve your issue? Offhand, I was more suspect of mysqld crashing - someone recently mentioned that holland is not taking innodb- log-files- in-group into account (and effectively defaulting to 2), for instance, which can lead to a crash of the snapshot mysqld instance. Perhaps there are other mysqld options not being considered correctly here.