Changing Enswitch UID and GID

Overview

It may be necessary to change the UID and GID of the enswitch user/group on an existing system. In our case the enswitch user was UID 100 and the enswitch group was GID 101, which caused us to have to renumber the existing 100 user and 101 group on every new server install.

Disclaimer

This is provided as-is with no warranty and may not work correctly on every Enswitch system. Make sure you have proper backups and test the procedure in a non-production environment. Neither I nor Integrics are responsible for any problems arising from the use of these intstructions.

Procedure

Variables

In this example, the enswitch UID is 100 and will be changed to 900, and the enswitch GID is 101 and will chaneg to 901.

Identify files owned by enswitch user and group

On each Enswitch server, find files owned by uid 100 or gid 101 and save them to a text file.

For NFS servers:

time find / -user 100 | egrep -v "\/proc|\/dev\/" > /tmp/enswitch_user_files.txt
time find / -group 101 | egrep -v "\/proc|\/dev\/" > /tmp/enswitch_group_files.txt

For all other servers:

time find / -user 100 | egrep -v "\/var\/lib\/enswitch\/|\/proc|\/dev\/" > /tmp/enswitch_user_files.txt
time find / -group 101 | egrep -v "\/var\/lib\/enswitch\/|\/proc|\/dev\/" > /tmp/enswitch_group_files.txt

Stop all services using enswitch user/group

Web (Active only):

/etc/init.d/apache2 stop

Asterisk:

/etc/init.d/asterisk stop
/etc/init.d/enswitch_routed stop
/etc/init.d/hylafax stop
chown -h enswitch:enswitch /var/spool/asterisk/voicemail
sleep 2
pgrep -lf asterisk
pgrep -lf enswitch_routed

Kamailio (Active only):

/etc/ha.d/resource.d/kamailio stop
/etc/ha.d/resource.d/rtpengine stop
/etc/ha.d/resource.d/enswitch_sipd stop
/etc/ha.d/resource.d/enswitch_messaged stop
sleep 2
pgrep -alf enswitch_sipd
pgrep -alf enswitch_messaged
pgrep -alf kamailio
pgrep -alf rtpengine

Change enswitch UID and GID

getent group enswitch
groupmod -g 901 enswitch
getent group enswitch

getent passwd enswitch
usermod -u 900 -g 901 enswitch
getent passwd enswitch

Change file ownership

Change ownership on all files that reside on the local disk on each server. Start this on the NFS boxes first because they will take the longest.

time cat /tmp/enswitch_user_files.txt | xargs --max-args=1000 chown enswitch
time cat /tmp/enswitch_group_files.txt | xargs --max-args=1000 chgrp enswitch

Once this has finished on all servers, move on to the next step

Start services

Web (Active only):

/etc/init.d/apache2 start

Asterisk:

/etc/init.d/asterisk start
/etc/init.d/enswitch_routed start
/etc/init.d/hylafax start

pgrep -lf asterisk
pgrep -lf enswitch_routed

Kamailio (Active only):

/etc/ha.d/resource.d/kamailio start
/etc/ha.d/resource.d/rtpengine start
/etc/ha.d/resource.d/enswitch_sipd start
/etc/ha.d/resource.d/enswitch_messaged start

pgrep -alf enswitch_sipd
pgrep -alf enswitch_messaged
pgrep -alf kamailio
pgrep -alf rtpengine

enswitch restart

Restart Enswitch on all other servers

enswitch restart

= Re-check ownership

On each Enswitch server, find files owned by uid 100 or gid 101 that may have been created after the initial check:

time find / -user 100 | egrep -v "\/var\/lib\/enswitch\/|\/proc|\/dev\/" > /tmp/enswitch_user_files_2.txt
time find / -group 101 | egrep -v "\/var\/lib\/enswitch\/|\/proc|\/dev\/" > /tmp/enswitch_group_files_2.txt

If any files are found, change ownership on them, then stop and start all services again.

time cat /tmp/enswitch_user_files_2.txt | xargs --max-args=1000 chown enswitch
time cat /tmp/enswitch_group_files_2.txt | xargs --max-args=1000 chgrp enswitch

Re-start cron

Restart cron on all boxes. I had an issue where enswitch_cdrs_archive, enswitch_cdrs_delete and other Enswitch cron jobs did not run after the change. Apparently cron caches the user to uid and group to gid mapping. Restarting cron fixed the issue.