ZoneMinder Forums

Posted: **Wed May 26, 2010 9:42 pm**

Hi Community!

I am not a linux expert so will try to explain the best I can:

I have a 1.23 setup with 16CH capture card. Setup was running with no problems until (I guess) HD filled up to 90% which is where it is at running Purgewhenfull, I mention this but, dunno if this could have something to do with the problem.

About every 5 minutes (havent clocked it precisely) TOP shows mysqld-max using 20% processor time, load goes above 2, during which a whole bunch of Buffer overrun at index xxx in all dvideo0, dvideo1, dvideo2 and dvideo3 WAR nings spit up on the messages.log

Monitoring TOP, once mysqld-max "finishes" or goes "away", load drops back under 1 and ZM is logging as normal, responsive and "fine and dandy".

During the time mysqld-max pushes the load on the server, sometimes events show up with no images (or broken images) I think because of the buffer overrun and the mess this error creates on newly created events.

I have up to now, deleted events in the HD as to show a 71% full thinking it had to do with Purgewhenfull deleting events phisically from the database and making everything run behind, but even with Purgewhenfull not working (because HD is under the stipulated 90%) I still get WAR Buffer overrun errors.

The 16 cams are in modect and produce in all about 200 motion detection events per hour, something I do not belive to be overload on the server. The server is a Pentium 4 machine with 2GB or ram and did handle this load without problems for months before.

The loss of realtime events while this Buffer overrun thing is what worries me.

Any light on how to debug this problem?

Regards!

Posted: **Thu May 27, 2010 4:16 am**

What types of drives are you running scsi ide sata?

Also how old are they ?

Sound like a i/o problem.

Look up how to do tests with the command hdpram to check speed.

Posted: **Thu May 27, 2010 3:04 pm**

Hi

I am running a single IDE drive, these are the results:

With ZM running, 3 tests:

Code: Select all

# hdparm -tT /dev/hda

/dev/hda:
 Timing cached reads:   2344 MB in  2.00 seconds = 1172.35 MB/sec
 Timing buffered disk reads:   72 MB in  3.20 seconds =  22.52 MB/sec
[root@sanl1v16 /]# hdparm -tT /dev/hda

/dev/hda:
 Timing cached reads:   2388 MB in  2.00 seconds = 1193.39 MB/sec
 Timing buffered disk reads:   96 MB in  3.01 seconds =  31.88 MB/sec
[root@sanl1v16 /]# hdparm -tT /dev/hda

/dev/hda:
 Timing cached reads:   2328 MB in  2.00 seconds = 1164.33 MB/sec
 Timing buffered disk reads:  108 MB in  3.00 seconds =  35.96 MB/sec

With ZM not running, 3 tests:

Code: Select all

# hdparm -tT /dev/hda

/dev/hda:
 Timing cached reads:   2592 MB in  2.00 seconds = 1296.81 MB/sec
 Timing buffered disk reads:  138 MB in  3.09 seconds =  44.73 MB/sec

/dev/hda:
 Timing cached reads:   2924 MB in  2.00 seconds = 1462.36 MB/sec
 Timing buffered disk reads:  158 MB in  3.02 seconds =  52.25 MB/sec

/dev/hda:
 Timing cached reads:   2856 MB in  2.00 seconds = 1427.15 MB/sec
 Timing buffered disk reads:  158 MB in  3.03 seconds =  52.12 MB/sec

With ZM running and WAR Bufferoverrun at index xxx

Code: Select all

# hdparm -tT /dev/hda

/dev/hda:
 Timing cached reads:   2344 MB in  2.00 seconds = 1172.60 MB/sec
 Timing buffered disk reads:  108 MB in  3.01 seconds =  35.92 MB/sec
[root@sanl1v16 /]# hdparm -tT /dev/hda

/dev/hda:
 Timing cached reads:   2380 MB in  2.00 seconds = 1190.30 MB/sec
 Timing buffered disk reads:  130 MB in  3.01 seconds =  43.13 MB/sec
[root@sanl1v16 /]# hdparm -tT /dev/hda

/dev/hda:
 Timing cached reads:   2468 MB in  2.00 seconds = 1236.53 MB/sec
 Timing buffered disk reads:  126 MB in  3.07 seconds =  41.01 MB/sec

Again when WAR Buffer overrun is on, mysqld-max TOPs with:

Code: Select all

 PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
 2717 mysql     15   0  140m  38m 4832 S 19.3  3.8 123:43.57 mysqld-max

What does mysql-max do?

Can its priority perhaps be changed?

Regards!

Posted: **Fri May 28, 2010 6:06 am**

If you know enofe mysql check to see if you have a crashed db OR the easy way phpmyadmin I recommended its very easy to see when you have -30000 records

ovesely thats wrong.

Posted: **Fri May 28, 2010 3:37 pm**

kingofkya

Sorry, did not understand your fix, how can I check to see what you are saying?

Thanks

Posted: **Sat May 29, 2010 5:52 am**

In phpMyAdmin theres a option to repair tables/databases you may have a "crashed" table where some of the data dosen't match up.

ZoneMinder Forums

mysqld-max and Buffer overrun at index xxx

mysqld-max and Buffer overrun at index xxx