renayama19661014 at ybb
Sep 20, 2011, 5:28 PM
Post #8 of 10
(461 views)
Permalink
|
|
Re: [Patch 3]Change avoiding the stop error of the mysql resource agent.
[In reply to]
|
|
Hi Raoul,
Thank you for comment.
> ok, but why do you ommit the check if the pidfile exists and then cat
> this very file. if you cat a non existing file, you'll get errors.
> basically, the ra does the following:
>
> > kill `cat /tmp/mysql.pid 2>/dev/null` > /dev/null; echo $?
> > kill: usage: kill [-s sigspec | -n signum | -sigspec] pid | jobspec ... or kill -l [sigspec]
> > 1
>
> so the ra will exit with:
> > ocf_log err "MySQL couldn't be stopped"
> > return $OCF_ERR_GENERIC
However, it is not necessary for the stop to become the error because mysql falls.
The error of the stop restrains FO in some cases.
* The similar processing to this problem is carried out in pgsql.
> shouldn't the following patch be enough
> (just adding the mysql_status check but not removing the pid check?)
>
> > diff --git a/heartbeat/mysql b/heartbeat/mysql
> > index e449de4..474f62e 100755
> > --- a/heartbeat/mysql
> > +++ b/heartbeat/mysql
> > @@ -898,6 +898,12 @@ mysql_stop() {
> > $CRM_MASTER -D
> > fi
> >
> > + mysql_status info
> > + rc=$?
> > + if [ $rc = $OCF_NOT_RUNNING ]; then
> > + return $OCF_SUCCESS
> > + fi
> > +
> > if [ ! -f $OCF_RESKEY_pid ]; then
> > ocf_log info "MySQL is not running"
> > return $OCF_SUCCESS
No.
Because it repeats with status processing, it should delete it that RA checks pid file.
- if [ ! -f $OCF_RESKEY_pid ]; then
- ocf_log info "MySQL is not running"
- return $OCF_SUCCESS
+ mysql_status info
+ rc=$?
+ if [ $rc = $OCF_NOT_RUNNING ]; then
+ return $OCF_SUCCESS
fi
Sorry...My understanding to your comment may be wrong.
Best Regards,
Hideo Yamauchi.
--- On Tue, 2011/9/20, Raoul Bhatia [IPAX] <r.bhatia [at] ipax> wrote:
> hello hideo-san!
>
> On 09/20/2011 02:19 AM, renayama19661014 [at] ybb wrote:
> ...
> >>> - if [ ! -f $OCF_RESKEY_pid ]; then
> >>> - ocf_log info "MySQL is not running"
> >>> - return $OCF_SUCCESS
> >>> + mysql_status info
> >>> + rc=$?
> >>> + if [ $rc = $OCF_NOT_RUNNING ]; then
> >>> + return $OCF_SUCCESS
> >>> fi
> ...
>
> >> i'm sorry but i do not understand the problem you're addressing.
> >> can you please describe with different words?
> >
> > Sorry....
> >
> > For example, mysql fails in stop processing and causes an error when the next trouble happens.
> >
> > Step1 ) For example, for switch over, stop handling of Mysql begins.
> > Step2 ) However, Mysql fell by process trouble just after that. The pid file is left.
> > Step3 ) When pid file is left by the current stop processing, an error happens.
> >
> > * The stop processing is finished normally by checking pid file and the existence of the process by this patch definitely.
> > * And the switch over excess succeeds.
>
> ok, but why do you ommit the check if the pidfile exists and then cat
> this very file. if you cat a non existing file, you'll get errors.
> basically, the ra does the following:
>
> > kill `cat /tmp/mysql.pid 2>/dev/null` > /dev/null; echo $?
> > kill: usage: kill [-s sigspec | -n signum | -sigspec] pid | jobspec ... or kill -l [sigspec]
> > 1
>
> so the ra will exit with:
> > ocf_log err "MySQL couldn't be stopped"
> > return $OCF_ERR_GENERIC
>
> shouldn't the following patch be enough
> (just adding the mysql_status check but not removing the pid check?)
>
> > diff --git a/heartbeat/mysql b/heartbeat/mysql
> > index e449de4..474f62e 100755
> > --- a/heartbeat/mysql
> > +++ b/heartbeat/mysql
> > @@ -898,6 +898,12 @@ mysql_stop() {
> > $CRM_MASTER -D
> > fi
> >
> > + mysql_status info
> > + rc=$?
> > + if [ $rc = $OCF_NOT_RUNNING ]; then
> > + return $OCF_SUCCESS
> > + fi
> > +
> > if [ ! -f $OCF_RESKEY_pid ]; then
> > ocf_log info "MySQL is not running"
> > return $OCF_SUCCESS
>
> cheers,
> raoul
>
_______________________________________________________
Linux-HA-Dev: Linux-HA-Dev [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
|
mysql.1651-3.patch
