【cluster with single master and single slave】dead-master unexcepted

[余国聪]

When a cluster has only single master and single slave,if master has a  network problem in  several seconds ,while the ReadTopologyInstanceBufferable function checked as a dead-master may cause a takeover we don't want.how can we avoid this problem?

Maybe setting a longer instancePollSecondsDuration can solve this problem,but actually not.

So can we check dead-master  repeatedly to ensure a deadmaster is really a deadmaster?

Thanks!

[Shlomi Noach]

> So can we check dead-master  repeatedly to ensure a deadmaster is really a deadmaster?

This is an approach orchestrator tries to keep away from; I don't know that there's a generic number of times and length of intervals to making the call.

You, can, however, use PreFailoverProcesses to run a `sleep 5 ; if [ your-own-master-is-now reachable ] ; then exit 1 ; fi`, does that make sense?

--
You received this message because you are subscribed to the Google Groups "orchestrator-mysql" group.
To unsubscribe from this group and stop receiving emails from it, send an email to orchestrator-my...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/orchestrator-mysql/dfe804b9-fd69-437c-9d5e-ef61bc2a9734o%40googlegroups.com.

[余国聪]

Thanks for your responds,using PreFailoverProcesses may avoid this problem most but not all the time.

Actually ,as you known, in  production environment,there may be many proxys and tools around master host, these tools may need to be changed when master Takeover,so the cost of master Taskover could be considered,which may give us a parameter to control to make the judge of dead-master more tolerant.

在 2020年6月16日星期二 UTC+8下午3:02:14，Shlomi Noach写道：

> So can we check dead-master  repeatedly to ensure a deadmaster is really a deadmaster?

This is an approach orchestrator tries to keep away from; I don't know that there's a generic number of times and length of intervals to making the call.

You, can, however, use PreFailoverProcesses to run a `sleep 5 ; if [ your-own-master-is-now reachable ] ; then exit 1 ; fi`, does that make sense?

On Tue, Jun 16, 2020 at 6:51 AM 余国聪 <gcy...@gmail.com> wrote:

When a cluster has only single master and single slave,if master has a  network problem in  several seconds ,while the ReadTopologyInstanceBufferable function checked as a dead-master may cause a takeover we don't want.how can we avoid this problem?

Maybe setting a longer instancePollSecondsDuration can solve this problem,but actually not.

So can we check dead-master  repeatedly to ensure a deadmaster is really a deadmaster?

Thanks!

--
You received this message because you are subscribed to the Google Groups "orchestrator-mysql" group.

To unsubscribe from this group and stop receiving emails from it, send an email to orchestrator-mysql+unsub...@googlegroups.com.