Problems with dataskq ...

O

OMNI

New member
Joined
Jul 23, 2011
Messages
1
Hi! I have a problem when you last updated direactadmin then appeared on many servers load caused dataskq.
Process hung and downloaded the server even before it.


OS CENTOS 5.6 (64)

4784 root 25 0 242m 187m 696 R 10.0 4.8 64:06.33 dataskq
13001 root 25 0 239m 185m 696 R 10.0 4.7 15:03.75 dataskq
20442 root 25 0 247m 192m 696 R 10.0 4.9 143:24.39 dataskq
25639 root 25 0 246m 191m 696 R 10.0 4.8 120:34.84 dataskq
27756 nobody 25 0 257m 200m 588 R 10.0 5.1 303:52.04 dataskq
31322 nobody 25 0 274m 216m 588 R 10.0 5.5 617:57.37 dataskq
16046 nobody 25 0 260m 203m 592 R 9.0 5.1 353:04.17 dataskq
1271 root 25 0 244m 188m 696 R 6.7 4.8 81:43.86 dataskq
16126 nobody 25 0 268m 210m 588 R 6.7 5.3 495:40.47 dataskq
30021 root 25 0 245m 189m 696 R 6.7 4.8 100:27.32 dataskq
16461 nobody 25 0 263m 206m 588 R 4.7 5.2 413:22.53 dataskq
9721 root 25 0 251m 195m 588 R 3.3 4.9 196:24.75 dataskq
16028 nobody 25 0 238m 184m 592 R 3.0 4.7 1:04.88 dataskq
31507 root 25 0 255m 198m 588 R 3.0 5.0 262:59.69 dataskq
9910 root 25 0 240m 186m 696 R 2.3 4.7 30:26.71 dataskq
Click to expand...

1. In the analysis of the log errortaskq.log can be seen:

2011:07:21-01:25:02: removing old lock: ./data/admin/brute.conf.lock (age: 10801 seconds)
2011:07:21-04:25:03: removing old lock: ./data/admin/brute.conf.lock (age: 10801 seconds)
2011:07:21-07:26:01: removing old lock: ./data/admin/brute.conf.lock (age: 10858 seconds)
2011:07:21-10:26:03: removing old lock: ./data/admin/brute.conf.lock (age: 10801 seconds)
2011:07:21-13:27:01: removing old lock: ./data/admin/brute.conf.lock (age: 10858 seconds)
2011:07:21-16:27:02: removing old lock: ./data/admin/brute.conf.lock (age: 10801 seconds)
2011:07:21-19:27:03: removing old lock: ./data/admin/brute.conf.lock (age: 10801 seconds)
2011:07:21-22:27:04: removing old lock: ./data/admin/brute.conf.lock (age: 10801 seconds)
2011:07:22-01:28:01: removing old lock: ./data/admin/brute.conf.lock (age: 10857 seconds)
2011:07:22-04:28:03: removing old lock: ./data/admin/brute.conf.lock (age: 10801 seconds)
2011:07:22-07:29:02: removing old lock: ./data/admin/brute.conf.lock (age: 10859 seconds)
2011:07:22-12:21:02: removing old lock: ./data/admin/brute.conf.lock (age: 10801 seconds)
2011:07:22-15:21:03: removing old lock: ./data/admin/brute.conf.lock (age: 10801 seconds)
2011:07:22-18:22:01: removing old lock: ./data/admin/brute.conf.lock (age: 10858 seconds)
Click to expand...

2. Somewhere there was a problem with:

2011:07:22-00:10:02: webalizer binary (/usr/bin/webalizer) is 0 bytes in size. Download a new binary: http://files.directadmin.com/services/
2011:07:22-00:10:02: webalizer binary (/usr/bin/webalizer) is 0 bytes in size. Download a new binary: http://files.directadmin.com/services/
2011:07:22-00:10:02: webalizer binary (/usr/bin/webalizer) is 0 bytes in size. Download a new binary: http://files.directadmin.com/services/
2011:07:22-00:10:02: webalizer binary (/usr/bin/webalizer) is 0 bytes in size. Download a new binary: http://files.directadmin.com/services/
2011:07:22-00:10:02: webalizer binary (/usr/bin/webalizer) is 0 bytes in size. Download a new binary: http://files.directadmin.com/services/
2011:07:22-00:10:02: webalizer binary (/usr/bin/webalizer) is 0 bytes in size. Download a new binary: http://files.directadmin.com/services/
2011:07:22-00:10:02: webalizer binary (/usr/bin/webalizer) is 0 bytes in size. Download a new binary: http://files.directadmin.com/services/
2011:07:22-00:10:02: webalizer binary (/usr/bin/webalizer) is 0 bytes in size. Download a new binary: http://files.directadmin.com/services/
2011:07:22-00:10:02: webalizer binary (/usr/bin/webalizer) is 0 bytes in size. Download a new binary: http://files.directadmin.com/services/
Click to expand...


When was the second problem, the process has not uploaded dataskq server. After fixing the second problem, the process is no longer hang dataskq

Maybe it's a bug with dataskq, if there is a problem now it hangs in the process.

How to fix the problem 1?
 
While I am not having the webalizer issue (I have it disabled anyway in directadmin.conf), I am receiving the brute force log entries like you. The catch is, my brute force option is disabled as well in directadmin.conf. And yet still my dataskq processes are also running out of control.

I instead use CSF for all brute force logging, and after disabling the native DA BF monitor, I restarted the directadmin service to no avail.

Does anyone else have any ideas why this is happening? My load is unbelievably high and gets higher and higher by the minute.... this has only started happening since the last couple of DA updates.

Thanks.
 
Last edited:
Load makes the process dataskq, so it is necessary to lower priority:
in file /etc/cron.d/directadmin_cron add to task dataskq priority, as example
Code: 
* * * * * root nice -n 2000 /usr/local/directadmin/dataskq
 
*** c u t ***

this is a temporary solution. I see your problem too, so we are waiting fixing from the developer.
 
Hello,

This is likely the cause:
http://www.directadmin.com/features.php?id=1243

The pre-release binaries are avialable. Click your license, and follow the pre-release link.

I'll also see about changing the logic, so entries are only appended to the end of the brute_log_entries.list, and the sort/removal of old values is only done once daily with the tally.

I'm hoping to get that logic finished in the next few days and will roll out a bugfix release.

John
 
I've just finished the new logic. It's exponentially more efficient.
They've just been uploaded. Ensure the "./directadmin o" output shows August 23rd to ensure you have the newest version.

John
 
You must log in or register to reply here.
Back
Top