0

I have a four server cluster with Oracle Clusterware managing 4 MySQL instances on them. On failure Clusterware will move the MySQL instance to another server.

I am on Proliant servers and also have ASR running. Recently I had a failover where ASR hard rebooted a server and Clusterware had to step in to bring up the database on another server.

I'm quite open to disabling ASR if necessary.

What production configurations should I run with ASR and Clusterware?

1 Answer 1

1

What was the cause for the ASR? Given that the watchdog timer caught the system crash, do you have anything in the HP IML log indicating the root problem?

I typically leave the ASR functionality on, even making a point to reduce the ASR timeout to 5 minutes from the default 10 minute setting. What is the desired outcome of your setup? An ASR indicates a larger problem with the OS/hardware. Are you saying that you want the server to stall and/or stay off so the other systems in the cluster take over? Or is it okay if the system reboots?

2
  • It is fine if the server reboots. I was thinking perhaps it would be easier if Oracle Clusterware does the failover before the machine fails as MySQL replication often fails on an ASR forced restart but not when Oracle moves the MySQL instance. I'll check the IML log. Jan 28, 2011 at 9:22
  • If that's the case, you can disable the ASR or maximize the ASR timeout (such that the Clusterware picks up the node failure first). But fix your hardware either way. ASRs should not be common.
    – ewwhite
    Jan 28, 2011 at 17:24

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .