Forum Discussion
Marius_Frankows
Nimbostratus
Nimbostratusdefine failover pool member
Hi all,
I need to define a 'failover' functionality via irule to a pool with two members.
Within the pool I do have one 'master' member and one 'slave' member. The master member is initially selected via priority group activation.
In case that this master server is marked as down the slave member will be reselect automatically because of the lower priority. This runs without any rule at all.
But now I need to preserve that the slave member remains selected even the master member is marked up.
Is there any way to set the priority group activation to the slave member via irule, so after master member fails once the slave member becomes higher priority and remains selected ?
The reason for this strange configuration is that these two members are mysql servers nodes in a master master replication and simply checking the port 3306 is not enough as healthcheck to validate the completion of the starting process after the mysl server failed or have been restarted.
I'd be apprecieted for any ideas on this topic
Thanks!
3 Replies
Matt_Breedlove_I would suggest you setup a cron job or scheduled task that collects the key metrics you want to make sure the SQL server is up and then have that script print out an HTML page to disk.
Serve that page over HTTP and point the ECV to that page.
That way your ECV doesnt muck with SQL on it's native port
We setup a system for mission critical nodes where there is a supervisor process that runs every minute and a few diagnostics that also run every minute. The diagnostics collect data and update a diag specific file in a directory that the supervisor then consumes and evaluates before overwriting the served html page.
Robert_Flach_11Can't you simply use the "manual resume" function?
I have never used it before, but it sounds like it is what you are looking for...
Marius_Frankowsafter my first tests the Click here described configuration works perfectly for my needs. Thanks for your support...
Marius
