0.80.2 upgrade results in infocom task getting stuck.

adrianmarsh · 2011-08-08 16:37:42

Hi,

Just updated my production system to 80.2

For some reason, the infocom task gets stuck into a running state.

GLPI seems to of created its own ticket stating:

Description : [GLPI] Monitoring of automatic actions
The following automatic actions are in error. They require intervention.

infocom : Send alarms on financial and administrative informations

I'm not sure what logging I need to turn on to figure this out. All worked ok on 0.78.2 before.

adrianmarsh · 2011-08-10 19:08:14

no-one else seeing this ? Or know what it means ?

Henchffs · 2011-08-15 15:49:55

I'm having a similar problem:

[GLPI] Monitoring of automatic actions

The following automatic actions are in error. They require intervention.
session : Clean expired sessions

Started popping up after upgrade to 0.80

Help is much appreciated.

Cheers!

adrianmarsh · 2011-09-19 13:33:54

I'm now getting several automated tickets a week generated by the helpdesk by infocom.

And I've still no idea:

a) what is it complaining about ?
b) how do I fix that
c) or how do I turn off the auto-ticket generation..

adrianmarsh · 2011-10-12 12:19:49

I'm still getting this... I've no way to disable 'infocom'.

I'll try an upgrade to 0.80.4

MoYo · 2011-10-18 09:48:36

You must go to Automatic action to clear execution in error. See logs to understand why there are troubles.
You could launch them manually to test.

If you do not want this automatic action you could disable them.

adrianmarsh · 2011-10-18 19:28:41

Hi Moyo,

Yes thats what I've been trying to do (view the logs), but I can't find anything complaining about infocom.

I just saw how to "disable" (under Status), but as I would like reminders, I'd like it to work.

Any tips on where to look for debug ?

Under /var/www/it/glpi/files/_log ??

I've looked through there for infocom and all I see are cron.log messages saying its launching, but nothing else.

MoYo · 2011-10-18 20:06:29

yes under /var/www/it/glpi/files/_log.
you could pass in debug mode too. Mayeb you will see informations online.
And launch manually the cron job.

adrianmarsh · 2011-10-19 15:16:32

Mmmm.. ok I set debug level to 5, then I cleared the log files in _log
Then I ran the job manually from the webpage.

It sat in the 'running' state for ages (not doing anything), and all that showed in the _log folder was one file:

cron.log

with:

19-10-2011 14:06 [39@ubiq-serv2]
Internal #1: Launch infocom
19-10-2011 14:06 [39@ubiq-serv2]
Internal #1: Nothing to launch

The state of the the job is still "Running" even now..

Nothing came up on the webpage display (actually after I hit 'execute' - literally nothing came up). I had to hit Refresh in the browser.

Below is the system config output (obscured a bit). it complains about the log folder, but its owned by apache:apache, and its creating log files in there, httpd process is run as apache.

For some reason the php-errors.log file has been recreated as root.

[root@ubiq-serv2 _log]# pwd
/var/www/it/glpi/files/_log
[root@ubiq-serv2 _log]# ls -la
total 20
drwxr-xr-x 3 apache apache 4096 Oct 19 14:13 .
drwxr-xr-x 27 apache apache 4096 Oct 19 14:12 ..
-rw-r--r-- 1 apache apache 609 Oct 19 14:13 cron.log
drwxr-xr-x 2 root root 4096 Oct 19 14:06 old
-rw-r--r-- 1 root root 1600 Oct 19 14:13 php-errors.log

Information about system installation and configuration

 
GLPI  0.80.2 (/glpi => /var/www/it/glpi)


Server

 
Operating Systems : Linux ubiq-serv2 2.6.18-238.5.1.el5 #1 SMP Mon Feb 21 05:52:39 EST 2011 x86_64
PHP 5.1.6 (PDO, Reflection, SPL, SimpleXML, apache2handler, bz2, calendar, ctype, curl, date, dbase, dom, exif, ftp, gd,
    gettext, gmp, hash, iconv, imap, json, ldap, libxml, mbstring, mime_magic, mysql, mysqli, openssl, pcre, pdo_mysql, pdo_sqlite,
    posix, pspell, session, shmop, sockets, standard, sysvmsg, sysvsem, sysvshm, tokenizer, wddx, xml, xmlreader, xmlwriter, xsl,
    zlib)
Setup: memory_limit="64M" max_execution_time="30" safe_mode="" session.save_handler="files" post_max_size="8M"
    upload_max_filesize="10M" 
Software: Apache/2.2.3 (Red Hat) (Apache/2.2.3 (Red Hat) Server at it.mycompany.local Port 443)
    Mozilla/5.0 (Windows NT 5.1; rv:6.0) Gecko/20100101 Firefox/6.0
MySQL: 5.0.77 (glpi@localhost/glpi)

../config : OK
../files : OK
../files/_dumps : OK
../files/_sessions : OK
../files/_cron : OK
../files/_cache/ : OK
../files/_graphs : OK
Check permissions to the directory: : ../files/_log


LDAP directory

 
Server: 'ldaps://vm-dc01.mycompany.local', Port (default=389): '636', Basedn: 'OU=MyBusiness,DC=mycompany,DC=local', Connection
        filter: '(&(objectClass=user)(objectCategory=person)(!(userAccountControl:1.2.840.113556.1.4.803:=2)))', rootdn (for non
        anonymous binds): 'CN=MYACCOUNT,OU=SBSUsers,OU=Users,OU=MyBusiness,DC=mycompany,DC=local', Use TLS: '0'


Mysql replicate

 
Active : No


Notifications / Mail Receiver

 
Way of sending emails: SMTP (helpdesk@smtp.mycompany.local)
Mail Receiver
    Name:"helpdesk@mycompany.local"   Server:{imap.mycompany.local/imap/novalidate-cert/notls} Login:"helpdesk" Password:Yes
        Active:Yes


OCS Inventory NG

 
Host for the OCSNG database : 'ubiq-serv1.mycompany.local', Connection to OCSNG database successful, Use the OCSNG software
        dictionary : 0


Plugins list

 
    datainjection    Name:File injection                 Version :2.1.0      Status :Enabled
    reports          Name:Reports                        Version :1.5.0      Status :Enabled

MoYo · 2011-10-25 08:23:20

php-errors must not be owned by root.
What is the data in this file ?

adrianmarsh · 2011-10-25 15:28:56

Hi,

ok, as part of my 0.80.2 -> 0.80.4 upgrade I reset the ownership of php-errors.log anyway, but, root ownership came by default.

Process-wise, root owns the top level httpd, but not the child-processes:

[root@ubiq-serv2 _log]# ps -ef|grep 24731
apache 3539 24731 0 Oct24 ? 00:00:15 /usr/sbin/httpd
apache 3540 24731 0 Oct24 ? 00:00:18 /usr/sbin/httpd
apache 3542 24731 0 Oct24 ? 00:00:15 /usr/sbin/httpd
apache 3543 24731 0 Oct24 ? 00:00:17 /usr/sbin/httpd
apache 3544 24731 0 Oct24 ? 00:00:14 /usr/sbin/httpd
apache 3546 24731 0 Oct24 ? 00:00:15 /usr/sbin/httpd
apache 3630 24731 0 Oct24 ? 00:00:15 /usr/sbin/httpd
apache 3676 24731 0 Oct24 ? 00:00:13 /usr/sbin/httpd
apache 3940 24731 0 Oct24 ? 00:00:13 /usr/sbin/httpd
apache 4431 24731 0 Oct24 ? 00:00:11 /usr/sbin/httpd
root 24731 1 0 Sep19 ? 00:00:05 /usr/sbin/httpd

All thats in that file now (since 0.80.4) is repeats of :

25-10-2011 14:16 [@ubiq-serv2]
Notice(8): Unknown: rsh to IMAP server timed out (errflg=1)
Backtrace :

I've re-run infocom, with debug at 0, and also at debug level 5, but nothing comes out into php-error.log

File ownerships are now:

-rw-r--r-- 1 apache apache 531511 Oct 25 14:27 cron.log
-rw-r--r-- 1 apache apache 42362 Oct 25 13:35 mail.log
drwxr-xr-x 2 apache apache 4096 Oct 19 14:06 old
-rw-r--r-- 1 apache apache 1492991 Oct 25 14:26 php-errors.log
-rw-r--r-- 1 apache apache 80 Sep 27 09:42 remove.txt

adrianmarsh · 2011-10-25 15:35:40

BTW: Server is running RHEL5, with php and httpd at:

httpd-manual-2.2.3-45.el5
httpd-2.2.3-45.el5

php-mbstring-5.1.6-27.el5_5.3
php-cli-5.1.6-27.el5_5.3
php-pear-1.4.9-6.el5
php-imap-5.1.6-27.el5_5.3
php-xml-5.1.6-27.el5_5.3
php-mysql-5.1.6-27.el5_5.3
php-common-5.1.6-27.el5_5.3
php-pdo-5.1.6-27.el5_5.3
php-ldap-5.1.6-27.el5_5.3
php-5.1.6-27.el5_5.3
php-pecl-json-1.2.1-4.el5
php-gd-5.1.6-27.el5_5.3
php-devel-5.1.6-27.el5_5.3

adrianmarsh · 2011-11-15 20:21:49

I still get this with infocom.

The log isn't owned by root, and it still stalls. Nothing more comes out in php-error.log if I set it to level 5.

I do get the word "Test" coming out into the log when I edit the General settings to change log level.

adrianmarsh · 2011-11-22 17:11:49

I've had to disable infocom, which is annoying as I wanted to use it.

I'm not getting any debug out either when I set debug to level 5, which is a little worrying.

adrianmarsh · 2011-11-22 17:15:05

Ah ha!!

I ran the cronjob manually, and I get the below error:

[root@ubiq-serv2 ~]# /usr/bin/php /var/www/it/glpi/front/cron.php
PHP Fatal error: Class '127' not found in /var/www/it/glpi/inc/infocom.class.php on line 392
[root@ubiq-serv2 ~]#

no idea what it means though :-(

MoYo · 2011-11-29 09:33:19

Class 127 may be an old Class index. Not used now. Maybe a plugin or something else.
You could delete all row in glpi_infocoms table whit such an itemtype

adrianmarsh · 2011-11-30 19:34:17

anyway of finding out?

When I run an sql query:
SELECT * FROM `glpi`.`glpi_infocoms` where itemtype="127"

I get 29 results... We have removed old plugins that didn't work with 0.80.x

Some of it looks like reminders for software purchases. None of it looks like plugin data though.

MoYo · 2011-11-30 21:58:11

So delete these lines.

solar · 2011-12-21 16:37:28

but this error I often gap.tuy course is a good indication of how far

Last edited by solar (2011-12-29 10:58:03)

solar · 2012-01-01 03:40:26

happy new year!

adrianmarsh · 2012-01-03 19:25:54

Ok - I've updated mine, deleting those lines out (after dumping them to a CSV in case I need them back again later). I've re-enabled infocom and will wait to see what happens. It has just sent me an email with about 50 expired warranty messages in it.

Forum GLPI-Project

Announcement

#1 2011-08-08 16:37:42

0.80.2 upgrade results in infocom task getting stuck.

#2 2011-08-10 19:08:14

Re: 0.80.2 upgrade results in infocom task getting stuck.

#3 2011-08-15 15:49:55

Re: 0.80.2 upgrade results in infocom task getting stuck.

#4 2011-09-19 13:33:54

Re: 0.80.2 upgrade results in infocom task getting stuck.

#5 2011-10-12 12:19:49

Re: 0.80.2 upgrade results in infocom task getting stuck.

#6 2011-10-18 09:48:36

Re: 0.80.2 upgrade results in infocom task getting stuck.

#7 2011-10-18 19:28:41

Re: 0.80.2 upgrade results in infocom task getting stuck.

#8 2011-10-18 20:06:29

Re: 0.80.2 upgrade results in infocom task getting stuck.

#9 2011-10-19 15:16:32

Re: 0.80.2 upgrade results in infocom task getting stuck.

#10 2011-10-25 08:23:20

Re: 0.80.2 upgrade results in infocom task getting stuck.

#11 2011-10-25 15:28:56

Re: 0.80.2 upgrade results in infocom task getting stuck.

#12 2011-10-25 15:35:40

Re: 0.80.2 upgrade results in infocom task getting stuck.

#13 2011-11-15 20:21:49

Re: 0.80.2 upgrade results in infocom task getting stuck.

#14 2011-11-22 17:11:49

Re: 0.80.2 upgrade results in infocom task getting stuck.

#15 2011-11-22 17:15:05

Re: 0.80.2 upgrade results in infocom task getting stuck.

#16 2011-11-29 09:33:19

Re: 0.80.2 upgrade results in infocom task getting stuck.

#17 2011-11-30 19:34:17

Re: 0.80.2 upgrade results in infocom task getting stuck.

#18 2011-11-30 21:58:11

Re: 0.80.2 upgrade results in infocom task getting stuck.

#19 2011-12-21 16:37:28

Re: 0.80.2 upgrade results in infocom task getting stuck.

#20 2012-01-01 03:40:26

Re: 0.80.2 upgrade results in infocom task getting stuck.

#21 2012-01-03 19:25:54

Re: 0.80.2 upgrade results in infocom task getting stuck.

Board footer