clustercheck hangs when server is hung
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Percona XtraDB Cluster moved to https://jira.percona.com/projects/PXC |
Fix Released
|
Undecided
|
Raghavendra D Prabhu |
Bug Description
When the mysql server on a node is hung clustercheck will also hang and never fail the server over.
Steps to reproduce:
1) setup a 3 node cluster
2) configure ha_proxy to run with an active writer and 2 backup servers
3) start a sysbench test
4) log into the writer node and run 'kill -SIGSTOP <mysqld_pid>' as root to hang the server
Setup details can be found here: http://
The sysbench test should timeout and the ha_proxy writer vip should also hang.
Attached is a python version of clustercheck that will timeout after 30 seconds and mark the node as failed. I didn't create a branch since this would add a new python dependency to PXC so, I wasn't sure if this needed to be written in perl or not in order to be accepted.
Related branches
Changed in percona-xtradb-cluster: | |
milestone: | none → 5.5.30-23.7.4 |
assignee: | nobody → Raghavendra D Prabhu (raghavendra-prabhu) |
status: | Confirmed → Fix Released |
Tested. The existing script can be reused by using timeout in front of mysql like http:// sprunge. us/KiRU
Tested the python script too, and it works.