Forum Discussion

bezeqint's avatar
bezeqint
Icon for Nimbostratus rankNimbostratus
Jan 14, 2013

IRule for pool member priority

hi all,

 

i have this issue with two pool members that one of the is the master and the other one is backup member.

 

 

the issue occur when the first one go down the backup become active and everything is working fiine, when the first one become active the BIG-IP doesn't close the connections to the backup.

 

is it possible to do by irule to close the connection as soon as the primary is active and the backup will go back to the old status.

 

 

Regards,

 

application delivery
dev
devops
iRules

8 Replies

  • Arie's avatar
    Arie
    Icon for Altostratus rankAltostratus
    Jan 14, 2013

    By default the LTM will gracefully handle existing connections so the user isn't affected. Is there a compelling reason to kick everyone off the backup server when the primary comes back online?

     

  • What_Lies_Bene1's avatar
    What_Lies_Bene1
    Icon for Cirrostratus rankCirrostratus
    Jan 14, 2013
    You could use an iRule but can I suggest you apply a OneConnect profile first? It should 'move' the connections more quickly although whether it's quick enough for you is another question.
  • bezeqint's avatar
    bezeqint
    Icon for Nimbostratus rankNimbostratus
    Jan 14, 2013
    hi,

     

    the issue is that the client open one connection and send the data only on one connection, so as soon as the primary pool member is going down the connection move to the backup server but as the primary server is coming up the connection remain on the backup sver, that causing problems with databases.
  • nitass's avatar
    nitass
    Icon for Employee rankEmployee
    Jan 14, 2013
    what if you delete existing connection (using bigpipe or tmsh) to backup server when primary server is up.

     

     

    Acton on Log - using the alertd deamon

     

    https://devcentral.f5.com/wiki/advdesignconfig.Acton-on-Log-using-the-alertd-deamon.ashx
  • bezeqint's avatar
    bezeqint
    Icon for Nimbostratus rankNimbostratus
    Jan 15, 2013
    that exactly what i would like to do.

     

    but is it possible to do this automatically?

     

     

    i don't want to do this manually, i heard that there is script that can handle this but i didn't understand the script at all.
  • Arie's avatar
    Arie
    Icon for Altostratus rankAltostratus
    Jan 16, 2013

    It sounds like the app wasn't developed with redundancy in mind. Any chance the developers could fix this? Or are you limited to using the LTM to work around the problem?

     

  • nitass's avatar
    nitass
    Icon for Employee rankEmployee
    Jan 17, 2013
    that exactly what i would like to do.

    but is it possible to do this automatically?

    i don't want to do this manually, i heard that there is script that can handle this but i didn't understand the script at all.e.g. 

    [root@ve10:Active] config  b virtual bar list
virtual bar {
   snat automap
   pool foo
   destination 172.28.19.252:23
   ip protocol 6
}
[root@ve10:Active] config  b pool foo list
pool foo {
   min active members 1
   monitor all tcp
   members {
      200.200.200.101:23 {
         priority 10
      }
      200.200.200.111:23 {
         priority 5
      }
   }
}
[root@ve10:Active] config  cat /config/user_alert.conf
alert BEZEQINT_TEST "Pool member 200.200.200.101:23 monitor status up." {
   exec command="export REMOTEUSER=root;/usr/bin/tmsh delete sys connection ss-server-addr 200.200.200.111"
}