[report] Plesk 10.4.4 statistics are never updated

[LaszloI]

ps aux | grep "cron"
root 7810 0.0 0.1 23848 1060 ? Ss Feb06 0:00 /usr/sbin/cron
root 26328 0.0 0.0 8968 864 pts/0 S+ 10:54 0:00 grep cron

In /etc/crontab :

# m h dom mon dow user command
3 * * * * root cd / && run-parts --report /etc/cron.hourly
6 1 * * * root test -x /usr/sbin/anacron || ( cd / && run-parts --report /etc/cron.daily )
35 1 * * 7 root test -x /usr/sbin/anacron || ( cd / && run-parts --report /etc/cron.weekly )
16 5 11 * * root test -x /usr/sbin/anacron || ( cd / && run-parts --report /etc/cron.monthly )
#

/usr/sbin/anacron file owner is root with permissions 755

/etc/anacrontab content:

SHELL=/bin/sh
PATH=/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin

# These replace cron's entries
1 5 cron.daily nice run-parts --report /etc/cron.daily
7 10 cron.weekly nice run-parts --report /etc/cron.weekly
@monthly 15 cron.monthly nice run-parts --report /etc/cron.monthly

 

[LaszloI]

I've started anacron with command:

start-stop-daemon --start --exec /usr/sbin/anacron -- -s

I hope this helps, I'll check it tomoroww morning.

 

[Webroy]

I had also problems with the statistics not updating

i did this:

mv /usr/local/psa/etc/logrotate.d/tomcat /root/logrotate_backup

and rebooted the server

after that it looks fine

 

[aashua]

Stats

Thanks for the update.
I will check and shall update this post.

 

[Pegasus2]

Plesk is up-to-date

Please install MU#5 http://kb.parallels.com/en/112937 where this issue was fixed.

I upgraded from 10.1.1 to 10.4.4 a few days ago.
Patch #18 is installed.

<product id="plesk" version="10.4.4" installed-at="20120219T154331">
<patch version="18" timestamp="" installed-at="20120219T154535" />
</product>

But I still cannot set the hosting type in a service plan.


Regards,
Harry

 

[BunyaminV]

Parallels Plesk Panel 10.x for Linux, 10.4.4 Update #19, Centos 6, x86_64

PROBLEM DESCRIPTION AND STEPS TO REPRODUCE
- The statistics not update daily.
- manually /usr/local/psa/admin/bin/statistics --calculate-all not works. waits long time then mysql server down.( i guess timeout)

- when i use --calculate-one -d parameter it updates but the statistics starts with that day and traffic is not calculated correctly.

ACTUAL RESULT
nothing updated daily

EXPECTED RESULT
daily update disk space and traffic.

 

[Faris Raouf]

Oddly, I'm being hit with this problem on a fresh install (MU23), with data migrated from another server.

Parallels Plesk Panel 10.x for Linux, 10.4.4 Update #23, Centos 6, x86_64

PROBLEM DESCRIPTION AND STEPS TO REPRODUCE
- Web statistics *and disk usage* not being updated automatically.
- Install 10.4.4 > check for updates > MU23 and all previous MUs installed > Migrate data from server running 8.6 > wait 24, 48 hours: No disk stats, no web stats.

NOTES:

/usr/local/psa/admin/sbin/statistics --calculate-one -domain-name=domain.tld works correctly and generates web usage stats and disk usage stats for domain.tld

/usr/local/psa/admin/sbin/statistics --calculate-one works correctly and generates web usage stats and disk usage stats for all domains.

** Unsupported Repo used: We have installed php 5.3.10 and MySQL 5.5 from ART repo. Also installed is ASL. But no errors in any logs relating to stats generation or control panel are seen.

** Unsupported Repo used: Repoforge repo is enabled for additional utilities such as multitail


ACTUAL RESULT
Disk and web stats not updated daily

EXPECTED RESULT
Disk and web stats updated daily

 

[IgorG]

What about Daily Maintenance script /etc/cron.daily/50plesk-daily with corresponding content

# install_statistics
/opt/psa/bin/sw-engine-pleskrun /opt/psa/admin/plib/DailyMaintainance/script.php >/dev/null 2>&1

? Does it started by cron? Did you try to run this command (psa/bin/sw-engine-pleskrun /opt/psa/admin/plib/DailyMaintainance/script.php >/dev/null 2>&1) manually?

 

[Faris Raouf]

(this post isn't really relevant. The cause is that nothing in cron.daily is being run at all in Centos 6 under Virtuozzo/OpenVZ template. Please see next post)

This is very interesting:

crond definitely running

/etc/cron.daily/50plesk-daily exists (root.root/-rwxr-xr-w) and contains

#!/bin/sh

# install_statistics
/usr/local/psa/bin/sw-engine-pleskrun /usr/local/psa/admin/plib/DailyMaintainance/script.php >/dev/null 2>&1

# install_mysqldump
/usr/local/psa/bin/mysqldump.sh >/dev/null 2>&1

As root, Manually running /usr/local/psa/bin/sw-engine-pleskrun /usr/local/psa/admin/plib/DailyMaintainance/script.php gives no errors and does update web-stats.

It also correctly updates disk stats

*** BUG?: After the script updates web-stats, it then goes on to run spamtrain / sa-learn. But I do not have psa-spamassassin installed. I do have spamassassin, but I have not installed the psa-specific functions via the installer. This wastes huge amounts of CPU for no good reason.

NOTE:

/etc/cron.daily also contains a file called awstats:

#!/bin/sh

/usr/share/awstats/awstats_updateall.pl now -awstatsprog=/var/www/cgi-bin/awstats/awstats.pl -configdir=/etc/awstats >/dev/null 2>&1

exit 0

Also
No mysql dumps (except pre-upgrade) are in /var/lib/psa/dumps -- I was expecting the other part of the plesk cron job to do that every day?

Conclusion: 50plesk-daily not actually being run by crond?

I have added an "echo Plesk script ran OK" to 50plesk-daily to see if this gets run at all daily.

 

[Faris Raouf]

OK. I have found the problem and the solution:

With reference to http://forum.openvz.org/index.php?t=msg&goto=44839&#msg_44839

In EL6, /etc/cron.hourly/0anacron is what triggers anacron (cronie-anacron), which is what actually causes daily cron jobs to run ( as defined in /etc/anacrontab )

Anyway, in the openvz forum post above, it points to a permissions issue in 0anacron.

And indeed, in he Virtuozzo Centos 6 template, 0anacron has the same (incorrect) permissions as referenced in that openvz post.

# ls -al /etc/cron.hourly/
total 20
drwxr-xr-x  2 root root 4096 Mar 21 17:29 .
drwxr-xr-x 85 root root 4096 Mar 26 13:58 ..
-rw-r--r--  1 root root  424 Dec 14 21:22 0anacron
-rwxr-xr-x  1 root root  630 Mar 19 18:35 asl
-rwxr-xr-x  1 root root  695 Mar  1  2010 freshclam

I'm surprised there's no error logged anywhere regarding this.

Because of this permission error, NOTHING in cron.daily gets run.

To fix it, just chmod 755 /etc/cron.hourly/0anacron

So if nothing in cron.daily (and therefore 50plesk-daily) does not appear to be running and you are running Centos 6 (or Scientific Linux 6) on Virtuozzo or OpenVZ via the appropriate OS template then it isn't a Plesk error - it is an error in the template. Just fix it as above.

Note to Igor - the spamtrain bug is still worth looking into please?: After the statistics script updates web-stats, it then goes on to run spamtrain (sa-learn) even if you do not have psa-spamassassin installed. This wastes huge amounts of CPU for no good reason.

 

[IgorG]

As far as I remember there were some optimizations for spamtraining procedure in MU#7...

 

[Alberto Tormo]

Check this thread, it solved my problem with stats.

 

[Faris Raouf]

As far as I remember there were some optimizations for spamtraining procedure in MU#7...

But why would they be active if psa-spamassassin is not installed? It doesn't seem like the correct behaviour.

Or, if I have misunderstood you, we have MU23 in this system, so if this was a bug corrected in MU7 then it doesn't appear to be corrected under EL6 64bit

 

[nocrash]

statistic not calculated

back to the important topic


---------------------------------------------------------------
PRODUCT, VERSION, OPERATING SYSTEM, ARCHITECTURE
Plesk 10.4.4 Update #25, 64 Bit Linux

/usr/local/psa/version:
10.4.4 Debian 6.0 1013111102.18

PROBLEM DESCRIPTION
statistics are not (maybe have been never) calculated since fresh one-click install at february this year

STEPS TO REPRODUCE
/usr/local/psa/admin/sbin/statistics -d XXX

results in "MySQL query failed: MySQL server has gone away" after a long time

ACTUAL RESULT
statistics are not calculated (0 byte traffic and webspace)

EXPECTED RESULT

ANY ADDITIONAL INFORMATION
ps fax shows that the statistics script executed awstats for every domain?

\_ /usr/bin/sw-engine -c /opt/psa/admin/conf/php.ini /usr/local/psa/admin/sbin/statistics -d XXX
\_ sh -c '/opt/psa/admin/bin/statistics_collector'
\_ /opt/psa/admin/bin/statistics_collector
\_ sh -c /usr/share/awstats/tools/awstats_buildstaticpages.pl -awstatsprog=/usr/lib/cgi-bin/awstats.pl -configdir=/opt/psa/etc/awstats -config=ANOTHERXXX-anon_ftp -dir=/var/www/vhosts/ANOTHERXXX/statistics/anon_ftpstat/current
\_ /usr/bin/perl /usr/share/awstats/tools/awstats_buildstaticpages.pl -awstatsprog=/usr/lib/cgi-bin/awstats.pl -configdir=/opt/psa/etc/awstats -config=ANOTHERXXX-anon_ftp -dir=/var/www/vhosts/ANOTHERXXX/statistics/anon_ftpstat/current
\_ sh -c "/usr/lib/cgi-bin/awstats.pl" -config=ANOTHERXXX-anon_ftp -staticlinks -configdir=/opt/psa/etc/awstats -output=allhosts 2>&1


--------------------------------------------------------------

 

[IgorG]

results in "MySQL query failed: MySQL server has gone away" after a long time

Looks like MySQL timeout. Did you try to increase wait_timeout in /etc/my.cnf:

wait_timeout 20000

and restart MySQL after that?

 

[nocrash]

sorry for the late reply , but thank you! this works..
he set the timouts to very low and high max connections before ..