Plesk 10 doesn't count disk usage and traffic, at all

[Renev]

We've recently spun up a new Plesk 10.x server in our Virtuozzo 4.6 cluster and have run into something strange.

For all clients the disk usage reads as '0'.
When I check the traffic usage screens under 'Tools & Utilities' everything reads 0 as well.

The only things that are counted are mailboxes and databases.

Could it be that there's some kind of cronjob that's not running or failing? Are there any log files that could provide more information?

The quotas work fine on our other VE's.

 

[IgorG]

Try to manually recalculate disk usage statistics for domain with

# /usr/local/psa/admin/sbin/statistics --calculate-one --domain-name=domain.tld

and check output.

 

[Renev]

Thanks. The first steps, HTTP and FTP stats seem to work fine, then it quits with this error:

error: error accessing /usr/share/tomcat6/logs: No such file or directory
error: tomcat:2 glob failed for /usr/share/tomcat6/logs/*.log

Naturally, Tomcat is not installed on this server. However, now the quotas do show up in Plesk for this domain!

 

[Matthew James]

Hi IgorG,

It seems if you use the --calculate-one it does update the Disk Usage in Plesk, however, it doesn't appear to when using --calculate-all (although it does do the AWStats update for all domains).

Also, the statistics_collector process seems to hang about:

root 19351 0.0 0.1 64848 3300 ? S 04:03 0:04 /usr/local/psa/admin/bin/statistics_collector

It's not doing anything:

#strace -p 19351
Process 19351 attached - interrupt to quit
wait4(13865,

Is this normal?

Thanks
Matt

 

[fogelf]

1. do not strace sbin/statistics - in not binary now.
if U will strace sbin/statistics process results will be not satisfied )

2. statistics corectly calculated only during daily mainteinance script
/usr/local/psa/bin/sw-engine-pleskrun /usr/local/psa/admin/plib/DailyMaintainance/script.php

 

[Matthew James]

Hi,

When is this script supposed to run, because statistics are not being calculated at all automatically.

When you try and do it manually, it just runs for HOURS, which was never a problem previously.

Thanks
Matt

 

[SALFATIM]

Hi,

When is this script supposed to run, because statistics are not being calculated at all automatically.

When you try and do it manually, it just runs for HOURS, which was never a problem previously.

Thanks
Matt

I confirm this point since I am on plesk 10.4.4. Statistics are never updated even if I launch the command manually
(http://forum.parallels.com/showthread.php?p=560290#post560290)

 

[JP Kelly]

Same Problem Here

I just noticed statistics have not been calculated for months!

Output of:
/usr/local/psa/admin/sbin/statistics --calculate-one --domain-name=mydomain.tld

WARNING during statistics_collector execution: : error: error accessing /usr/share/tomcat5/logs: No such file or directory
error: tomcat:2 glob failed for /usr/share/tomcat5/logs/*.log

Help!

 

[IgorG]

Just go to Home>Tools & Settings> Server settings and uncheck "Java applications" in "Include in the disk space usage calculation" if you haven't installed Tomcat.

 

[VVasilis]

Hello,

We are having the same issue with our server (10.4 migrated from 8.6).

I've disabled the Java Application in "Include in the disk space usage calculation" , but i receive the tomcat6 logs no such file error if i execute manually /usr/local/psa/admin/sbin/statistics --calculate-one --domain-name=domainltd . Is this expected behavior?

The upside is the stats are now shown for domainltd.

 

[fogelf]

Yes, it expected. Actually this utilitu is not for public execution. But if U exec it by Urself be ready to see some warns.

 

[JP Kelly]

/usr/local/psa/admin/sbin/statistics --calculate-all FAILS

/usr/local/psa/admin/sbin/statistics --calculate-all

fails with:

MySQL query failed: MySQL server has gone away

I have increased MySQL wait_timeout to wait_timeout = 3600 and I still have the problem.

 

[IgorG]

/usr/local/psa/admin/sbin/statistics --calculate-all

fails with:

MySQL query failed: MySQL server has gone away

I have increased MySQL wait_timeout to wait_timeout = 3600 and I still have the problem.

wait_timeout was 20000 here http://forum.parallels.com/showthread.php?t=208854&page=5 and it was successful.

 

[JP Kelly]

wait_timeout was 20000 here http://forum.parallels.com/showthread.php?t=208854&page=5 and it was successful.

Ok Ill give a try.

Does it always take 5+ hours to perform Stats update?

 

[IgorG]

Does it always take 5+ hours to perform Stats update?

It depends on number of sites and their popularity in Internet.

 

[fogelf]

>It depends on number of sites and their popularity in Internet.

Actually it just minor part of problem. Main goal here - is the webstat backend for site. By result of my tests (1000 domains with 1000 HTTP requests per domain) - if U use awstats calculation time is too long, webalizer is more fast. Minutes/seconds for webalizer (C backend) and hours for awstats(Perl backend).

If U want fast calculation - change statistics for all domains (simply shell-script). Results will be very nice.

 

[Alexey_Shpak]

Hello.

/usr/local/psa/admin/sbin/statistics --calculate-all

thereis no output

when I can see it log?
or can i change log to my file?

 

[fogelf]

In 11x sbin/statistics - is the PHP wrapper over sbin/statistics_collector. Wrapper is hides statistics_collector output.

 

[JP Kelly]

In 11x sbin/statistics - is the PHP wrapper over sbin/statistics_collector. Wrapper is hides statistics_collector output.

I'm not sure what this means.
Can you explain a bit more? Does it help with the problem of the statistics command timing out?

 

[fogelf]

look:
[xxx@yyy.com]# ls /usr/local/psa/admin/sbin | grep statistics
statistics - php script (wrapper over statistic_collector) - it call statistic_collector, parse and analyze its output
statistics_collector - binary for direct calculation, produces output, do not called directly
[xxx@yyy.com]#