Page 1 of 1

System Statistics Graph gone haywire

Posted: 14 Nov 2016, 19:01
by CatchThePigeon
My system statistics over the last 7 days shows a load spike of 4 million% and during the same period 60 trillion(!) apache connections...

[Connections Min:1.78 Max:6027691307339.00 Avg:8502157079.04]

This is certainly because cPanel auto-backup hit a snag and stopped deleting my rolling server backups. After a while this meant that I had completely run out of space on the server disks and things started to go wrong quickly.

On getting everything back to normal I've since logged in and checked my System Statistics page and the graph is showing an average load spike off the scale. Same with many of the other graphs.

I can't seem to find much information about where these logs are taken from or where they're stored but I'd like to remove the erroneous result as I find the interface quite useful to show. Also I don't like the idea of potentially corrupted log data remaining on the server as I've had problems with log corruptions causing a whole headache of other issues before.

If anyone with any ideas on the subject would be kind enough to reply I'd be very grateful.

Re: System Statistics Graph gone haywire

Posted: 14 Nov 2016, 20:57
by ForumAdmin
The data for the graphs is stored in csv format in:

Code: Select all

/var/lib/csf/stats/system
From lfd the stats line is recorded as:

Code: Select all

		print SYSSTAT "$time,$cputotal,$cpuidle,$cpuiowait,$memtotal,$memfree,$memswaptotal,$memswapfree,$load[0],$load[1],$load[2],$netin,$netout,$diskread,$diskwrite,$mailin,$mailout,$cputemp,$mysqlin,$mysqlout,$mysqlq,$mysqlsq,$mysqlcn,$mysqlth,$apachecpu,$apacheacc,$apachebwork,$apacheiwork,$diskw,$memcached\n";
which should give you information to establish what the columns are.

Re: System Statistics Graph gone haywire

Posted: 14 Nov 2016, 21:47
by CatchThePigeon
Thanks for that - it's really helpful.

I seem to have found the lines responsible and what's happened is that there's a few instances where it's writing the new line starting from unix time 147xxxxxxx in the middle of some results. So you may have one result that shown correctly as;
1476557993,5387685276,5164489378,9918109,16304404,1843320,8224764,8047068,0.04,0.04,0.00,35678222126,395237610139,32759306,86371571,0,0,,,,,,,,.00344678,897017,2,8,,11737672
Then another is:
1478753108,222530746,214705432,1477277,16304176,613016,8224764,8104791478755089,223321293,215489701,1477489,16304176,607464,8224764,8104832,0.00,0.00,0.00,1319619712,20469618215,2942374,3436510,0,0,,,,,,,,.0159109,359026,1,9,,13017428
Where the next is written in the middle of the last.

I edited out those lines and everything's back to normal.

Thanks for your help