mq.sh script

4 messages in this thread

list Neil Simmonds · Mon, 27 Jul 2020 15:39:56 +0100 ·

Hi all, I'm seeing some really strange behaviour from the MQ.SH script/

We have the standard mq.sh script that comes with Xymon and we have it
setup in /xymon/client/etc/clientlaunch.cfg as follows

[mq]
        ENVFILE $XYMONCLIENTHOME/etc/xymonclient.cfg
        CMD $XYMONCLIENTHOME/ext/mq.sh queuemanagername
        LOGFILE $XYMONCLIENTLOGS/xymonmq2client.log
        INTERVAL 1m

The behaviour that we are seeing is that,as expected MQ.SH writes nothing
in the way of errors into xymonmq2client.log but what it does write into
there is the contents of /xymon/client/tmp/logfetch.hostname.cfg

(i.e.
log:/applogs/broker/broker.log:20480
ignore MARK
)

I'm at a loss to explain this so am reaching out to the list to see if
anyone has come across this

Regards,
Neil.

list Damien Martins · Mon, 27 Jul 2020 18:44:19 +0200 ·

Hello,

I'm not using this script, but you may run it using:

bash -x? $XYMONCLIENTHOME/ext/mq.sh queuemanagername

replace $XYMONCLIENTHOME with the correct path.

Bash will generate output debug info, so you may see what it is doing.

Also, you may check if the log destination is hardcoded in the script.


Le 27/07/2020 ? 16:39, Neil Simmonds a ?crit?:

▸ quoted from Neil Simmonds

Hi all, I'm seeing some really strange behaviour from the MQ.SH script/

We have the standard mq.sh script that comes with Xymon and we have it 
setup in /xymon/client/etc/clientlaunch.cfg as follows

[mq]
? ? ? ? ENVFILE $XYMONCLIENTHOME/etc/xymonclient.cfg
? ? ? ? CMD $XYMONCLIENTHOME/ext/mq.sh queuemanagername
? ? ? ? LOGFILE $XYMONCLIENTLOGS/xymonmq2client.log
? ? ? ? INTERVAL 1m

The behaviour that we are seeing is that,as expected MQ.SH writes 
nothing in the way of errors into xymonmq2client.log but what it does 
write into there is the contents of 
/xymon/client/tmp/logfetch.hostname.cfg

(i.e.
log:/applogs/broker/broker.log:20480
ignore MARK
)

I'm at a loss to explain this so am reaching out to the list to see if 
anyone has come across this

Regards,
Neil.

list Neil Simmonds · Tue, 28 Jul 2020 14:58:03 +0100 ·

I've already tried running it with a set -x in the script and get the
following in the log,

+ rm /xymon/client/tmp/mq-server_name.11677
+ TMPFILE=/xymon/client/tmp/mq-server_name.11684
+ touch /xymon/client/tmp/mq-server_name.11684
+ echo 'client server_name.mqcollect mqcollect'
+ '[' 1 '!=' 0 ']'
+ QMGR=PPIM01
+ shift
+ sudo -u mqm mqsc PPIM01 'display qstatus(*) curdepth msgage lgetdate
lgettime'
+ '[' 0 '!=' 0 ']'
+ /xymon/client/bin/xymon ukclamon01 @
log:/opt/IBM/MDMPIM/logs/workflowengine_server_name/db_manager.log:10240
log:`find -L /opt/IBM/MDMPIM/logs/\w*server_name/\w*.log -maxdepth 1  -type
f ! -size 0`:10240
log:/applogs/mdmpim/logs/appsvr_server_name/default.log:10240
log:/opt/IBM/MDMPIM/logs/workflowengine_server_name/customegl.log:20480
log:/opt/IBM/MDMPIM/logs/workflowengine_server_name/default.log:20480
ignore INFO
trigger internal error
file:`find /applogs/mdmpim/dump -name *.txt`
file:`find /applogs/mdmpim/dump -name *.phd``

From there I can see that it appears to be the line that sends the message
to Xymon that's doing it but U can't see why.

The line in question is underlined and BOLD below (copied the whole script
for context)


TMPFILE="$XYMONTMP/mq-$MACHINE.$$"

touch $TMPFILE

echo "client $MACHINE.mqcollect mqcollect" >$TMPFILE

#
while [ $# != 0 ]
do
    QMGR=$1; shift
    sudo -u mqm mqsc $QMGR 'display qstatus(*) curdepth msgage lgetdate
lgettime' >> $TMPFILE
    ##(echo 'dis ql(*) curdepth'; echo 'dis chs(*)'; echo 'end') | su - mqm
-c "runmqsc $QMGR" >> $TMPFILE
done
#

*$XYMON $XYMSRV "@" < $TMPFILE*
rm $TMPFILE

exit 0


On Mon, Jul 27, 2020 at 5:54 PM Damien Martins via Xymon <xymon at xymon.com>
wrote:

---------- Forwarded message ----------
From: Damien Martins <user-c12727b399f0@xymon.invalid>
To: xymon at xymon.com
Cc:
Bcc:
Date: Mon, 27 Jul 2020 18:44:19 +0200
Subject: Re: [Xymon] mq.sh script

▸ quoted from Damien Martins


Hello,

I'm not using this script, but you may run it using:

bash -x  $XYMONCLIENTHOME/ext/mq.sh queuemanagername

replace $XYMONCLIENTHOME with the correct path.

Bash will generate output debug info, so you may see what it is doing.

Also, you may check if the log destination is hardcoded in the script.


Le 27/07/2020 ? 16:39, Neil Simmonds a ?crit :

Hi all, I'm seeing some really strange behaviour from the MQ.SH script/

We have the standard mq.sh script that comes with Xymon and we have it
setup in /xymon/client/etc/clientlaunch.cfg as follows

[mq]
        ENVFILE $XYMONCLIENTHOME/etc/xymonclient.cfg
        CMD $XYMONCLIENTHOME/ext/mq.sh queuemanagername
        LOGFILE $XYMONCLIENTLOGS/xymonmq2client.log
        INTERVAL 1m

The behaviour that we are seeing is that,as expected MQ.SH writes nothing
in the way of errors into xymonmq2client.log but what it does write into
there is the contents of /xymon/client/tmp/logfetch.hostname.cfg

(i.e.
log:/applogs/broker/broker.log:20480
ignore MARK
)

I'm at a loss to explain this so am reaching out to the list to see if
anyone has come across this

Regards,
Neil.


Xymon mailing user-d459c9d661b6@xymon.invalid


---------- Forwarded message ----------
From: Damien Martins via Xymon <xymon at xymon.com>
To: xymon at xymon.com
Cc:
Bcc:
Date: Mon, 27 Jul 2020 18:44:19 +0200
Subject: Re: [Xymon] mq.sh script

list Jeremy Laidman · Wed, 29 Jul 2020 15:19:34 +1000 ·

Neil

I can confirm that I can replicate this behaviour when sending a xymon
"client" message using the command structure in the script, and it is
returning the relevant section of client-local.cfg. This is the expected
and documented behaviour of the "xymon client" command - see man(5)
client-local.cfg for info about how this works with the client message.

Interestingly. your script is using the "testname is mqcollect" way to
specify a collector ID. Newer versions of the mq.sh script doesn't rely on
this special behaviour, and instead sets the collector ID by suffixing the
"client message type with the collector ID, using this line:

echo "client/mqcollect $MACHINE.mqcollect mqcollect" >$TMPFILE

You seem to be using a script that's likely more than 10 years old. I
recommend upgrading if you can.

And I suspect this is the crux of the problem, in that the script was
written when xymond didn't give a response to a xymon client command. That
was a long time ago.

Although I doubt the age of the mq.sh script is causing any issues for you,
and certainly shouldn't be the cause of the inappropriate lines in the
logfile. My copy of mq.sh, from Xymon 4.3.10, has the same problem.

A solution to this is for this script to use "-" instead of "@" in the
xymon command, and in doing so, have the received output discarded (it's
still received, but xymon doesn't present it on STDOUT), like so:

$XYMON $XYMSRV "-" < $TMPFILE

Another option would be to simply redirect STDOUT to /dev/null:

$XYMON $XYMSRV "@" < $TMPFILE >/dev/null

This will still log anything that is sent to STDERR. As we would not expect
any output, unless there was a problem, dumping all standard output is
unlikely to have any negative consequences. Personally, I prefer the "-"
solution.

Cheers
Jeremy

On Tue, 28 Jul 2020 at 23:58, Neil Simmonds <user-feff97fabd3d@xymon.invalid>

▸ quoted from Neil Simmonds

wrote:

I've already tried running it with a set -x in the script and get the
following in the log,

+ rm /xymon/client/tmp/mq-server_name.11677
+ TMPFILE=/xymon/client/tmp/mq-server_name.11684
+ touch /xymon/client/tmp/mq-server_name.11684
+ echo 'client server_name.mqcollect mqcollect'
+ '[' 1 '!=' 0 ']'
+ QMGR=PPIM01
+ shift
+ sudo -u mqm mqsc PPIM01 'display qstatus(*) curdepth msgage lgetdate
lgettime'
+ '[' 0 '!=' 0 ']'
+ /xymon/client/bin/xymon ukclamon01 @
log:/opt/IBM/MDMPIM/logs/workflowengine_server_name/db_manager.log:10240
log:`find -L /opt/IBM/MDMPIM/logs/\w*server_name/\w*.log -maxdepth 1
-type f ! -size 0`:10240
log:/applogs/mdmpim/logs/appsvr_server_name/default.log:10240
log:/opt/IBM/MDMPIM/logs/workflowengine_server_name/customegl.log:20480
log:/opt/IBM/MDMPIM/logs/workflowengine_server_name/default.log:20480
ignore INFO
trigger internal error
file:`find /applogs/mdmpim/dump -name *.txt`
file:`find /applogs/mdmpim/dump -name *.phd``

From there I can see that it appears to be the line that sends the message
to Xymon that's doing it but U can't see why.

The line in question is underlined and BOLD below (copied the whole script
for context)


TMPFILE="$XYMONTMP/mq-$MACHINE.$$"

touch $TMPFILE

echo "client $MACHINE.mqcollect mqcollect" >$TMPFILE

#
while [ $# != 0 ]
do
    QMGR=$1; shift
    sudo -u mqm mqsc $QMGR 'display qstatus(*) curdepth msgage lgetdate
lgettime' >> $TMPFILE
    ##(echo 'dis ql(*) curdepth'; echo 'dis chs(*)'; echo 'end') | su -
mqm -c "runmqsc $QMGR" >> $TMPFILE
done
#

*$XYMON $XYMSRV "@" < $TMPFILE*
rm $TMPFILE

exit 0


On Mon, Jul 27, 2020 at 5:54 PM Damien Martins via Xymon <xymon at xymon.com>
wrote:

---------- Forwarded message ----------
From: Damien Martins <user-c12727b399f0@xymon.invalid>
To: xymon at xymon.com
Cc:
Bcc:
Date: Mon, 27 Jul 2020 18:44:19 +0200
Subject: Re: [Xymon] mq.sh script

Hello,

I'm not using this script, but you may run it using:

bash -x  $XYMONCLIENTHOME/ext/mq.sh queuemanagername

replace $XYMONCLIENTHOME with the correct path.

Bash will generate output debug info, so you may see what it is doing.

Also, you may check if the log destination is hardcoded in the script.


Le 27/07/2020 ? 16:39, Neil Simmonds a ?crit :

Hi all, I'm seeing some really strange behaviour from the MQ.SH script/

We have the standard mq.sh script that comes with Xymon and we have it
setup in /xymon/client/etc/clientlaunch.cfg as follows

[mq]
        ENVFILE $XYMONCLIENTHOME/etc/xymonclient.cfg
        CMD $XYMONCLIENTHOME/ext/mq.sh queuemanagername
        LOGFILE $XYMONCLIENTLOGS/xymonmq2client.log
        INTERVAL 1m

The behaviour that we are seeing is that,as expected MQ.SH writes nothing
in the way of errors into xymonmq2client.log but what it does write into
there is the contents of /xymon/client/tmp/logfetch.hostname.cfg

(i.e.
log:/applogs/broker/broker.log:20480
ignore MARK
)

I'm at a loss to explain this so am reaching out to the list to see if
anyone has come across this

Regards,
Neil.

Xymon mailing user-d459c9d661b6@xymon.invalid


---------- Forwarded message ----------
From: Damien Martins via Xymon <xymon at xymon.com>
To: xymon at xymon.com
Cc:
Bcc:
Date: Mon, 27 Jul 2020 18:44:19 +0200
Subject: Re: [Xymon] mq.sh script

mq.sh script 🔗 link

mq.sh script