MQ

Hi!

I'm trying to make the MQ Manager autostart after the AWS Instance restart or reboot. 

Basically trying to run a bash script when the server start but nothing happens...

#!/bin/bash
# Start queue managers
qmgrs=( $(dspmq | grep -v "Running" | cut -d "(" -f2 | cut -d ")" -f1) )
for i in "${qmgrs[@]}"
do
:
strmqm "$i"
done



Thank you!

Hi!

I'm trying to make the MQ Manager autostart after the AWS Instance restart or reboot. 

Basically trying to run a bash script when the server start but nothing happens...

#!/bin/bash
# Start queue managers
qmgrs=( $(dspmq | grep -v "Running" | cut -d "(" -f2 | cut -d ")" -f1) )
for i in "${qmgrs[@]}"
do
:
strmqm "$i"
done



Thank you!

Hi!

I'm trying to make the MQ Manager autostart after the AWS Instance restart or reboot. 

Basically trying to run a bash script when the server start but nothing happens...

#!/bin/bash
# Start queue managers
qmgrs=( $(dspmq | grep -v "Running" | cut -d "(" -f2 | cut -d ")" -f1) )
for i in "${qmgrs[@]}"
do
:
strmqm "$i"
done



Thank you!

------------------------------
Juan Reforme
------------------------------

You can run as a systemd service, example:

[Unit]Description=IBM MQ queue manager {QueueManagerName}Wants=network-online.targetAfter=network.target[Service]ExecStart=/opt/mqm/bin/strmqm -x {QueueManagerName}ExecStop=/opt/mqm/bin/endmqm -w -s {QueueManagerName}SuccessExitStatus=0 5 30 77Type=forkingUser=mqmGroup=mqmKillMode=noneLimitNOFILE=10240LimitNPROC=4096[Install]WantedBy=multi-user.target"

------------------------------
Matheus Francisco

Original Message:
Sent: Thu January 11, 2024 06:20 PM
From: Juan Reforme
Subject: MQ manager autostart after AWS reboot

Hi!

I'm trying to make the MQ Manager autostart after the AWS Instance restart or reboot. 

Basically trying to run a bash script when the server start but nothing happens...

#!/bin/bash
# Start queue managers
qmgrs=( $(dspmq | grep -v "Running" | cut -d "(" -f2 | cut -d ")" -f1) )
for i in "${qmgrs[@]}"
do
:
strmqm "$i"
done



Thank you!

------------------------------
Juan Reforme
------------------------------

Hi!

I'm trying to make the MQ Manager autostart after the AWS Instance restart or reboot. 

Basically trying to run a bash script when the server start but nothing happens...

#!/bin/bash
# Start queue managers
qmgrs=( $(dspmq | grep -v "Running" | cut -d "(" -f2 | cut -d ")" -f1) )
for i in "${qmgrs[@]}"
do
:
strmqm "$i"
done



Thank you!

Need code that works?

#!/bin/bash
for qm in $(dspmq | grep -v Running | cut -d'(' -f2 | cut -d')' -f1) ; do
strmqm $qm
done

Preferred solution for Linux is to use systemd.

Hi!

I'm trying to make the MQ Manager autostart after the AWS Instance restart or reboot. 

Basically trying to run a bash script when the server start but nothing happens...

#!/bin/bash
# Start queue managers
qmgrs=( $(dspmq | grep -v "Running" | cut -d "(" -f2 | cut -d ")" -f1) )
for i in "${qmgrs[@]}"
do
:
strmqm "$i"
done



Thank you!

regarding using systemd to start/stop a MQ Queue Manager, is it better to have systemd call the strmqm / endmqm commands directly, or better to have systemd call a ksh script that then calls startmqm/endmqm and whatever else the script may do?

Need code that works?

#!/bin/bash
for qm in $(dspmq | grep -v Running | cut -d'(' -f2 | cut -d')' -f1) ; do
strmqm $qm
done

Preferred solution for Linux is to use systemd.

Hi!

I'm trying to make the MQ Manager autostart after the AWS Instance restart or reboot. 

Basically trying to run a bash script when the server start but nothing happens...

#!/bin/bash
# Start queue managers
qmgrs=( $(dspmq | grep -v "Running" | cut -d "(" -f2 | cut -d ")" -f1) )
for i in "${qmgrs[@]}"
do
:
strmqm "$i"
done



Thank you!

I would always go via a script.

That way you can add things like your own logging, perhaps you need to do setmqenv or run from non-default directories, maybe do extra work during shutdown like forcibly killing apps that won't go away on their own ...

regarding using systemd to start/stop a MQ Queue Manager, is it better to have systemd call the strmqm / endmqm commands directly, or better to have systemd call a ksh script that then calls startmqm/endmqm and whatever else the script may do?

Need code that works?

#!/bin/bash
for qm in $(dspmq | grep -v Running | cut -d'(' -f2 | cut -d')' -f1) ; do
strmqm $qm
done

Preferred solution for Linux is to use systemd.

Hi!

I'm trying to make the MQ Manager autostart after the AWS Instance restart or reboot. 

Basically trying to run a bash script when the server start but nothing happens...

#!/bin/bash
# Start queue managers
qmgrs=( $(dspmq | grep -v "Running" | cut -d "(" -f2 | cut -d ")" -f1) )
for i in "${qmgrs[@]}"
do
:
strmqm "$i"
done



Thank you!

1. Don't even try to setup MQ systemd as a "user" service unless you really want it to start when the user logs on, and stop when the user exits. .  Its an effort in futility.  
2. When starting MQ from systemd in addition to using a script it is important to setup the ulimits that typically come with a login, but are not provided by systemd since systemd doesnt 'logon':

e.g.  mq.service 

[Unit]Description=systemd file to start MQWants=centrifydc.serviceAfter=centrifydc.service[Service]# Note in this example the actual location of the script is being purposefully obfuscated # to prevent discovery of infrastructure.ExecStart=/usr/bin/bash -c '<targetdir>/mqnace startmq'ExecStop=/usr/bin/bash -c '<targetdir>/mqnace stopmq'User=mqmType=forkingKillMode=noneLimitAS=infinityLimitCORE=infinityLimitDATA=infinityLimitFSIZE=infinityLimitMEMLOCK=65535LimitMSGQUEUE=819200LimitNOFILE=1048576LimitNPROC=1048576LimitRSS=infinityLimitSIGPENDING=192441LimitSTACK=8192000TimeoutStartSec=3600TimeoutStopSec=240[Install]WantedBy=multi-user.target

3. On our hosts we use a single script that derives the queue manager and broker info from the hostname and handles both the start and stop of mq AND/OR ACE as appropriate.

4. Endmqm may not in some cases be sufficient: there are some processes that dont stop nicely,  so the script needs to do something like 

# Brutally Killing any remaining mq processes"          
            for PROCESS in "${MQ_PROCESSES}"
            do # Output of kill redirected to /dev/null in case no processes
                ps -ef | grep "$PROCESS" | grep -v grep | \
                awk '{print $2}'| xargs kill -9
            done

I would always go via a script.

That way you can add things like your own logging, perhaps you need to do setmqenv or run from non-default directories, maybe do extra work during shutdown like forcibly killing apps that won't go away on their own ...

regarding using systemd to start/stop a MQ Queue Manager, is it better to have systemd call the strmqm / endmqm commands directly, or better to have systemd call a ksh script that then calls startmqm/endmqm and whatever else the script may do?

Need code that works?

#!/bin/bash
for qm in $(dspmq | grep -v Running | cut -d'(' -f2 | cut -d')' -f1) ; do
strmqm $qm
done

Preferred solution for Linux is to use systemd.

Hi!

I'm trying to make the MQ Manager autostart after the AWS Instance restart or reboot. 

Basically trying to run a bash script when the server start but nothing happens...

#!/bin/bash
# Start queue managers
qmgrs=( $(dspmq | grep -v "Running" | cut -d "(" -f2 | cut -d ")" -f1) )
for i in "${qmgrs[@]}"
do
:
strmqm "$i"
done



Thank you!

The script that is called by systemd to start the queue manager, couldn't it do something like this to accomplish the "logon" to pick up mqm's ulimits, env variables, etc

runuser -l  mqm -c "/opt/mqm/bin/strmqm $QMGR" 

1. Don't even try to setup MQ systemd as a "user" service unless you really want it to start when the user logs on, and stop when the user exits. .  Its an effort in futility.  
2. When starting MQ from systemd in addition to using a script it is important to setup the ulimits that typically come with a login, but are not provided by systemd since systemd doesnt 'logon':

e.g.  mq.service 

[Unit]Description=systemd file to start MQWants=centrifydc.serviceAfter=centrifydc.service[Service]# Note in this example the actual location of the script is being purposefully obfuscated # to prevent discovery of infrastructure.ExecStart=/usr/bin/bash -c '<targetdir>/mqnace startmq'ExecStop=/usr/bin/bash -c '<targetdir>/mqnace stopmq'User=mqmType=forkingKillMode=noneLimitAS=infinityLimitCORE=infinityLimitDATA=infinityLimitFSIZE=infinityLimitMEMLOCK=65535LimitMSGQUEUE=819200LimitNOFILE=1048576LimitNPROC=1048576LimitRSS=infinityLimitSIGPENDING=192441LimitSTACK=8192000TimeoutStartSec=3600TimeoutStopSec=240[Install]WantedBy=multi-user.target

3. On our hosts we use a single script that derives the queue manager and broker info from the hostname and handles both the start and stop of mq AND/OR ACE as appropriate.

4. Endmqm may not in some cases be sufficient: there are some processes that dont stop nicely,  so the script needs to do something like 

# Brutally Killing any remaining mq processes"          
            for PROCESS in "${MQ_PROCESSES}"
            do # Output of kill redirected to /dev/null in case no processes
                ps -ef | grep "$PROCESS" | grep -v grep | \
                awk '{print $2}'| xargs kill -9
            done

I would always go via a script.

That way you can add things like your own logging, perhaps you need to do setmqenv or run from non-default directories, maybe do extra work during shutdown like forcibly killing apps that won't go away on their own ...

regarding using systemd to start/stop a MQ Queue Manager, is it better to have systemd call the strmqm / endmqm commands directly, or better to have systemd call a ksh script that then calls startmqm/endmqm and whatever else the script may do?

Need code that works?

#!/bin/bash
for qm in $(dspmq | grep -v Running | cut -d'(' -f2 | cut -d')' -f1) ; do
strmqm $qm
done

Preferred solution for Linux is to use systemd.

Hi!

I'm trying to make the MQ Manager autostart after the AWS Instance restart or reboot. 

Basically trying to run a bash script when the server start but nothing happens...

#!/bin/bash
# Start queue managers
qmgrs=( $(dspmq | grep -v "Running" | cut -d "(" -f2 | cut -d ")" -f1) )
for i in "${qmgrs[@]}"
do
:
strmqm "$i"
done



Thank you!

This probably would work.  In our case most of our queue managers run on one - and only one of a pair:  therefore we needed logic to figure out which host was designated active in the pair.  

For configuration management reasons, we wanted to have a single script that would execute on any of our 100+ mq hosts handling both start and stop. In addition, as I remarked under some circumstances endmqm does not stop all the mqm processes on the host so scripting the stop was a requirement.

We chose also to have the same script handle the start/stop ace which simplifies the configuration management even more.

We experimented with systemd templates , but the management  was more convoluted and prone to error.

The script that is called by systemd to start the queue manager, couldn't it do something like this to accomplish the "logon" to pick up mqm's ulimits, env variables, etc

runuser -l  mqm -c "/opt/mqm/bin/strmqm $QMGR" 

1. Don't even try to setup MQ systemd as a "user" service unless you really want it to start when the user logs on, and stop when the user exits. .  Its an effort in futility.  
2. When starting MQ from systemd in addition to using a script it is important to setup the ulimits that typically come with a login, but are not provided by systemd since systemd doesnt 'logon':

e.g.  mq.service 

[Unit]Description=systemd file to start MQWants=centrifydc.serviceAfter=centrifydc.service[Service]# Note in this example the actual location of the script is being purposefully obfuscated # to prevent discovery of infrastructure.ExecStart=/usr/bin/bash -c '<targetdir>/mqnace startmq'ExecStop=/usr/bin/bash -c '<targetdir>/mqnace stopmq'User=mqmType=forkingKillMode=noneLimitAS=infinityLimitCORE=infinityLimitDATA=infinityLimitFSIZE=infinityLimitMEMLOCK=65535LimitMSGQUEUE=819200LimitNOFILE=1048576LimitNPROC=1048576LimitRSS=infinityLimitSIGPENDING=192441LimitSTACK=8192000TimeoutStartSec=3600TimeoutStopSec=240[Install]WantedBy=multi-user.target

3. On our hosts we use a single script that derives the queue manager and broker info from the hostname and handles both the start and stop of mq AND/OR ACE as appropriate.

4. Endmqm may not in some cases be sufficient: there are some processes that dont stop nicely,  so the script needs to do something like 

# Brutally Killing any remaining mq processes"          
            for PROCESS in "${MQ_PROCESSES}"
            do # Output of kill redirected to /dev/null in case no processes
                ps -ef | grep "$PROCESS" | grep -v grep | \
                awk '{print $2}'| xargs kill -9
            done

I would always go via a script.

That way you can add things like your own logging, perhaps you need to do setmqenv or run from non-default directories, maybe do extra work during shutdown like forcibly killing apps that won't go away on their own ...

regarding using systemd to start/stop a MQ Queue Manager, is it better to have systemd call the strmqm / endmqm commands directly, or better to have systemd call a ksh script that then calls startmqm/endmqm and whatever else the script may do?

Need code that works?

#!/bin/bash
for qm in $(dspmq | grep -v Running | cut -d'(' -f2 | cut -d')' -f1) ; do
strmqm $qm
done

Preferred solution for Linux is to use systemd.

Hi!

I'm trying to make the MQ Manager autostart after the AWS Instance restart or reboot. 

Basically trying to run a bash script when the server start but nothing happens...

#!/bin/bash
# Start queue managers
qmgrs=( $(dspmq | grep -v "Running" | cut -d "(" -f2 | cut -d ")" -f1) )
for i in "${qmgrs[@]}"
do
:
strmqm "$i"
done



Thank you!

How do we prove to ourselves that the queue manager is in fact running with the system variable we think its running with?  I noticed at the bottom of an FDC file the queue manager shows a lot of that. Is there a way to cause the queue manager to eject a harmless info only FDC that we can then look at?

This probably would work.  In our case most of our queue managers run on one - and only one of a pair:  therefore we needed logic to figure out which host was designated active in the pair.  

For configuration management reasons, we wanted to have a single script that would execute on any of our 100+ mq hosts handling both start and stop. In addition, as I remarked under some circumstances endmqm does not stop all the mqm processes on the host so scripting the stop was a requirement.

We chose also to have the same script handle the start/stop ace which simplifies the configuration management even more.

We experimented with systemd templates , but the management  was more convoluted and prone to error.

The script that is called by systemd to start the queue manager, couldn't it do something like this to accomplish the "logon" to pick up mqm's ulimits, env variables, etc

runuser -l  mqm -c "/opt/mqm/bin/strmqm $QMGR" 

1. Don't even try to setup MQ systemd as a "user" service unless you really want it to start when the user logs on, and stop when the user exits. .  Its an effort in futility.  
2. When starting MQ from systemd in addition to using a script it is important to setup the ulimits that typically come with a login, but are not provided by systemd since systemd doesnt 'logon':

e.g.  mq.service 

[Unit]Description=systemd file to start MQWants=centrifydc.serviceAfter=centrifydc.service[Service]# Note in this example the actual location of the script is being purposefully obfuscated # to prevent discovery of infrastructure.ExecStart=/usr/bin/bash -c '<targetdir>/mqnace startmq'ExecStop=/usr/bin/bash -c '<targetdir>/mqnace stopmq'User=mqmType=forkingKillMode=noneLimitAS=infinityLimitCORE=infinityLimitDATA=infinityLimitFSIZE=infinityLimitMEMLOCK=65535LimitMSGQUEUE=819200LimitNOFILE=1048576LimitNPROC=1048576LimitRSS=infinityLimitSIGPENDING=192441LimitSTACK=8192000TimeoutStartSec=3600TimeoutStopSec=240[Install]WantedBy=multi-user.target

3. On our hosts we use a single script that derives the queue manager and broker info from the hostname and handles both the start and stop of mq AND/OR ACE as appropriate.

4. Endmqm may not in some cases be sufficient: there are some processes that dont stop nicely,  so the script needs to do something like 

# Brutally Killing any remaining mq processes"          
            for PROCESS in "${MQ_PROCESSES}"
            do # Output of kill redirected to /dev/null in case no processes
                ps -ef | grep "$PROCESS" | grep -v grep | \
                awk '{print $2}'| xargs kill -9
            done

I would always go via a script.

That way you can add things like your own logging, perhaps you need to do setmqenv or run from non-default directories, maybe do extra work during shutdown like forcibly killing apps that won't go away on their own ...

regarding using systemd to start/stop a MQ Queue Manager, is it better to have systemd call the strmqm / endmqm commands directly, or better to have systemd call a ksh script that then calls startmqm/endmqm and whatever else the script may do?

Need code that works?

#!/bin/bash
for qm in $(dspmq | grep -v Running | cut -d'(' -f2 | cut -d')' -f1) ; do
strmqm $qm
done

Preferred solution for Linux is to use systemd.

Hi!

I'm trying to make the MQ Manager autostart after the AWS Instance restart or reboot. 

Basically trying to run a bash script when the server start but nothing happens...

#!/bin/bash
# Start queue managers
qmgrs=( $(dspmq | grep -v "Running" | cut -d "(" -f2 | cut -d ")" -f1) )
for i in "${qmgrs[@]}"
do
:
strmqm "$i"
done



Thank you!