Open OliverKellyATech opened 10 years ago
show your /etc/xinetd.d/mysqlchk
This is happening for me too.
For anyone else having this issue, I was able to fix it by changing the mysqlchk xinetd script to use the root user.
This is happening for me too. Anyone can slove it ?
I have moved to another utility that is doing exactly the same, but with less effort (written in go, no xinetd required) https://github.com/leoleovich/clusterhc
Same problem here, even with the Percona-provided clustercheck script which is based on this here. Attempts like changing the xinetd user and making sure execute permissions are set on the MySQL socket directory didn't help.
Edit: Fixed by not quoting the password in the xinetd definition (in server_args
):
Fails:
[...]
server = /usr/bin/clustercheck
server_args = clustercheck "the-password"
[...]
Works:
[...]
server = /usr/bin/clustercheck
server_args = clustercheck the-password
[...]
Related issue https://bugzilla.redhat.com/show_bug.cgi?id=459553
And for use credentials from .cnf
file as workaround use the wrapper script /usr/bin/clustercheck.wrapper
like:
#!/bin/sh
/usr/bin/clustercheck "" "" $@
And xinetd config should be:
server = /usr/bin/clustercheck.wrapper
server_args = 0 /dev/null 0 /etc/mysqlchk.cnf
Dont amend chmod +x /usr/bin/clustercheck.wrapper
Hi guys,
Quick question; I am getting a mismatch between the clustercheck response (via cli) and the response returned via the mysqlchk xinetd request:
/etc/xinetd.d/mysqlchk
is configured to simply callserver=/usr/local/bin/clustercheck
, as per the install instructions.Any ideas?