Admin Statistics

[bellzerr]

Statistics csv files ARE being created on my system, but the statistics charts never change. Some are empty, some display only a short, initial period of time. What should I check to get my statistical displays working correctly?
Thanks!

[soxfan]

Not sure if you're in the same boat as I am, but I posted concerning a similar issue last week. It seems that when I upgraded to 4.5.6 a Zimbra MySQL table became corrupted. Everything else besides the Zimbra statistics seems to be working fine, but my attempts to repair the table have failed. If you are having the same problem, and you come up with a solution please post back.

For your reference here's the link to my post, with details of where I found the info about the corrupt table, and what has been suggested to fix it.

[bellzerr]

Actually, I don't think my statistics ever worked.

[bellzerr]

Upgrading to 4.5.6 had the opposite effect for me - it FIXED my statistics. Everything except the DISK tab is working perfectly (it simply shows "Data are not yet available"). And the stats WERE being accumulated - just not properly displayed on the admin console. I now have stats going back a full year.

[easytouch]

I fixed today my logger and graphs with the following steps:

Read the Zimbra Wiki - Logger article and check that logger is working as expected
Logger - ZimbraWiki

check the contents of this directory
/opt/zimbra/zimbramon/rrdtool/work/

Verify that the gifs are created, and not of zero size
open one and verify that it's got a graph in it?
source: Disk: Statistics not produced

All commands beyond this point are run as user zimbra
# su zimbra

Is your hostname (as listed in the admin console) listed in the db?
# logmysql zimbra_logger
select distinct(host) from mta_aggregate;

+----------------------+
| host                 |
+----------------------+
| mx.examplea.com |
+----------------------+
1 row in set (0.01 sec)

exit (to leave mysql admin)
source: No data in server statistics


Check Disk Statistics in Zimbra's mySQL Database
logmysql zimbra_logger
select * from disk_aggregate;
select * from disk_status;
exit (to leave mysql admin)
do you get any rows displayed?


run logger manually
zimbra@mx:$ ~/libexec/zmlogprocess
(Attention! if lots are unprocessed, run it off business hours- cpu load,)
On a box with 200 mailboxes, 4000mails per day and 4 months of logs to process it took around half an hour.

check the running zmlogprocess from a second SSH shell
top
[Ctrl]+P (sort by CPU usage)

top - 11:47:42 up 2 days, 1:06, 3 users, load average: 2.25, 1.19, 0.52
Tasks: 125 total, 2 running, 123 sleeping, 0 stopped, 0 zombie
Cpu(s): 27.9% us, 4.6% sy, 0.0% ni, 36.3% id, 30.8% wa, 0.0% hi, 0.3% si
Mem: 3613204k total, 3492948k used, 120256k free, 21340k buffers
Swap: 8393952k total, 28k used, 8393924k free, 2258716k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
19855 zimbra 25 0 1338m 489m 16m S 0.0 13.9 0:51.05 java
18355 zimbra 16 0 1263m 234m 4300 S 0.0 6.6 0:16.12 mysqld
25587 zimbra 16 0 89776 84m 2060 R 30.3 2.4 0:27.55 zmlogprocess <=====

generate graphs by hand
# ~/libexec/zmgengraphs

after doing this, my graphs showed up.

I am not sure at the moment if they are updated by the cronjobs. Thats another issue I have to check later.

[grant@dcdata.co.za]

Hi guys,

My graphs done have any data displayed on them, the graphs are both there for server statistics and domain specific statistics but they just not displaying anything.

I have run through the checklist as per above and everything checks out except that above it refers to a folder, /opt/zimbra/zimbramon/rrdtool/work/
I dont have the work folder, in that directory I have the following folders.

bin contrib doc examples html include lib man var

Im not sure if this is just a version specific thing and might have changed. Im running 4.5.9, Im using the Suse 10 installation file but Im running SLES 10 SP1.

In the /opt/zimbra/zimbramon directory there is already a link for the rrdtool folder, rrdtool -> /opt/zimbra/zimbramon/rrdtool-1.0.49 , this was already there I didnt add it.

The only work folder that I have is in /opt/zimbra/zimbramon/mrtg/ but it has nothing inside it.

Any help would be greatly appreciated.

[dwmtractor]

When I had a similar problem it turned out that the logger daemon syslogd (which is basic to Linux, not to Zimbra) was not running. Once we restarted syslogd everything populated. You might check for that process.

[grant@dcdata.co.za]

Thanks for the tip. The server is running SLES 10 so it doesnt use the syslogd process, it uses a similar one called syslog-ng which is running and has been running since I made the changes yesterday.

I checked this server now and miraculously the graphs are now showing data. Perhaps it just takes a while to gather enough info before it can start logging. I was working on this problem yesterday morning and there was no graphs at 4pm, this morning at 9:13 am there are graphs and the graphs have logs going back to 12noon yesterday so I think it just needed about 24 hours of data before it could create the graphs.
Perhaps a cron job ran last night ran and fixed the graphs, I doubt it though as I tried manually creating the graphs yesterday and it didnt fix the problem.

[multicast]

When I run the command:
select distinct(host) from mta_aggregate;
I get "Empty" as a response....

Any ideas on how to fix that?