upgrade 6.0RC2 to 6.0.0. failed. now problems with bav.sock

weirdgod · Post by **weirdgod** » Thu Dec 03, 2009 12:58 am

After failed upgrade (i use mysql password), i tried to manually complete the upgrade (by mysqladmin -u root -p openemm < update_openemm-6.RC2-6.0.sql)... but this resulted some error (duplicate entry or something)...

but anyhow, when i stopped/run openemm again, it seems to work, but the only problem is the following:

-bash-3.2$ ./OpenEMM.sh stop
Stopping httpd
Stop scripts/slrtscn.py program .. delaying 2 seconds .. done.
Stop all sendmail processes: -15:18360 -15:18363 -15:18366 -15:18369 done.
Stop scripts/pickdist program .. delaying 2 seconds .. done.
Stop scripts/update.py program .. delaying 2 seconds .. done.
Stop scripts/bavd program .. delaying 2 seconds .. done.
Stop scripts/bav-trigger program with signal 15 .. done.
Stop scripts/bav-update program with signal 15 .. done.

-bash-3.2$ ./OpenEMM.sh start
Start /home/openemm/bin/scripts/bav-update.py .. done.
Start /home/openemm/bin/scripts/bav-trigger.py .. done.
Start /home/openemm/bin/scripts/bavd.py .. done.
Start /home/openemm/bin/bav -L INFO .. done.
Start /home/openemm/bin/scripts/update.py bounce account .. done.
Start /home/openemm/bin/scripts/pickdist.py .. done.
Stopping obsolete sendmail processes: done.
Starting sendmails: listener WARNING: Xbav: local socket name /home/openemm/var/run/bav.sock missing
client queue clientmqueue admin queue mail queue done.
Start /home/openemm/bin/scripts/slrtscn.py .. done.
Resin httpd start at Thu Dec 3 00:57:29 CET 2009
-bash-3.2$

somehow bav does not run correctly, and i do not know why.

any suggestions?

thanks,
Jaka

Post by **maschoff** » Thu Dec 03, 2009 5:51 pm

You should NOT have used

Code: Select all

 mysqladmin -u root -p openemm < update_openemm-6.RC2-6.0.sql)

BUT

Code: Select all

mysql -u root -p openemm_cms < /usr/share/doc/OpenEMM-6.0/openemm_cms.sql

(taken from UPDATE.txt)

Your new installation seems to be incomplete. I recommend to backup emm.properties and make a fresh installation of OpenEMM manually. All importent data are in the database and file emm.properties, so nothing will be lost.

weirdgod · Post by **weirdgod** » Fri Dec 04, 2009 8:19 pm

I know.

but on my system there is no:
/usr/share/doc/OpenEMM-6.0 directory.
only OpenEMM-5.5.1 (which was the version i first used for full install, since then i upgraded automatically to 6.0RC2, and now to 6.0 final)...

so... ? what now, shall i download the full 6.0 package manually, extract it and try this command with openemm_cms.sql ? Are there any chances this might work?

Post by **maschoff** » Fri Dec 04, 2009 9:56 pm

As I have written before: I recommend to backup emm.properties and make a fresh installation of OpenEMM manually. You do not need a directory /usr/share/doc/OpenEMM-6.0 to do this.

weirdgod · Post by **weirdgod** » Sun Dec 13, 2009 1:37 pm

OK, following your suggestions i did the following:

1. logined as openemm, and run ./OpenEMM.sh stop
2. backed up openemm database
3. moved /home/openemm to /home/openemm.old
4. created new directory openemm, set permissions user/group to openemm
5. downloaded OpenEMM-6.0-bin.tar.gz from sourceforce, extracted it in home dir
6. copied emm.properties from old installation to new one
7. ran OpenEMM.sh start

i still get the same errors:
...
-bash-3.2$ ./OpenEMM.sh start
Start /home/openemm/bin/scripts/bav-update.py .. done.
Start /home/openemm/bin/scripts/bav-trigger.py .. done.
Start /home/openemm/bin/scripts/bavd.py .. done.
Start /home/openemm/bin/bav -L INFO .. done.
Start /home/openemm/bin/scripts/update.py bounce account .. done.
Start /home/openemm/bin/scripts/pickdist.py .. done.
Stopping obsolete sendmail processes: done.
Starting sendmails: listener 451 4.0.0 /etc/mail/sendmail.cf: line 1705: Xgreylist: local socket name /var/milter-greylist/milter-greylist.sock unsafe: Permission denied
WARNING: Xbav: local socket name /home/openemm/var/run/bav.sock missing
550 Permission denied (real uid not trusted)
client queue clientmqueue Permission denied (real uid not trusted)
admin queue WARNING: Ignoring submission mode -N option (not in submission mode)
Permission denied (real uid not trusted)
mail queue WARNING: Ignoring submission mode -N option (not in submission mode)
Permission denied (real uid not trusted)
done.
Start /home/openemm/bin/scripts/slrtscn.py .. done.
-bash-3.2$ Resin httpd start at Sun Dec 13 13:34:57 CET 2009
----

[root@sytech ~]# service sendmail start
Starting sendmail: WARNING: Xbav: local socket name /home/openemm/var/run/bav.sock missing
[ OK ]
Starting sm-client: [ OK ]

----

So, now it seems i have some persmission problems... and secondly - the missing bav.sock - indicates that bav is not running... what to do?

-Jaka

Post by **maschoff** » Mon Dec 14, 2009 5:16 pm

Re your step 5: Did you use tar option "p" for unpacking?

The milter-greylist error message indicates that sendmail is somehow configured in a special way. When in doubt remove the sendmail package and deploy it again to get a fresh sendmail installation.

weirdgod · Post by **weirdgod** » Tue Jan 05, 2010 11:42 pm

I found the problem...

I used the "p" (actually tar zxvfp OpenEMM-6.0.1.tar.gz)... but i did that as user openemm... which was the mistake, as files that were supposed to become setuid root, did not.

I just followed your suggestion and did a fresh install. All the settings/data remained in database, so now everything seems to work fine.

thanks!

Post by **maschoff** » Wed Jan 06, 2010 8:39 am

Thanks for your feedback!