[CWB] mysql cache errors occurring recently

[Hardie, Andrew]

If all tables are properly tracked, then the oldest should be deleted whenever more space is needed. So the cache will rise to 99/100 % full, then stay at that level of fullness. 

(Before a new DB is created, a check is made for the cache being past its limit. The " DATABASE CACHE OVERLOAD " message appears if that process fails for any reason.)

It's difficult for me to work out what the problem might be without actually being able to see your cache control window myself. 

One possibility is if your users have lots of categorised queries? These take up space in the DB cache, but cannot be deleted. If this is the problem, the solution is to grow the cache.

You can look at the list of databases in cache for each corpus in the admin tools of that corpus - under "Cached databases" . 

best

Andrew. 

-----Original Message-----
From: cwb-bounces at sslmit.unibo.it [mailto:cwb-bounces at sslmit.unibo.it] On Behalf Of Jörg Knappen
Sent: 12 April 2018 16:41
To: Open source development of the Corpus WorkBench <cwb at sslmit.unibo.it>
Subject: Re: [CWB] mysql cache errors occurring recently

Andrew,

thanks for the answer. I found the tool and there were indeed a few  
orphaned cache entries,
but their total size was negligible. The cache is still 99% in use;  
how can I clear more of it?

Jörg Knappen

Quoting "Hardie, Andrew" <a.hardie at lancaster.ac.uk>:

> 3 things.
>
> First, the function names/line numbers in the debug backtrace  
> suggest this is rather an old version of the code. So, you might  
> consider upgrading.
>
> Second, you might find that if you look on the "cache control" /  
> "database cache" page in the admin interface, it will tell you what  
> is wrong. (This is assuming that your version of the code is not too  
> old to have that tool!) One possibility is that you have a number of  
> orphaned database table records which make the cache look fuller  
> than it actually is. Deleting the orphan records would in that case  
> solve the problem.
>
> Third, when you get the " Too many database processes " error, you  
> can look under "Manage MySQL processes" again in the admin UI, and  
> delete records of zombie processes.
>
> best
>
> Andrew.
>
>
> -----Original Message-----
> From: cwb-bounces at sslmit.unibo.it  
> [mailto:cwb-bounces at sslmit.unibo.it] On Behalf Of Jörg Knappen
> Sent: 12 April 2018 13:48
> To: cwb at sslmit.unibo.it
> Subject: [CWB] mysql cache errors occurring recently
>
> We run a cqpweb on Ubuntu 14.04 LTS using the following php version
>
>
> PHP 5.5.9-1ubuntu4.24 (cli) (built: Mar 16 2018 12:41:43)
> Copyright (c) 1997-2014 The PHP Group
> Zend Engine v2.5.0, Copyright (c) 1998-2014 Zend Technologies
>      with Zend OPcache v7.0.3, Copyright (c) 1999-2014, by Zend Technologies
>
>
> and the mysql version
>
> mysql  Ver 14.14 Distrib 5.5.59, for debian-linux-gnu (x86_64) using
> readline 6.3
>
> We did not change cqpweb recently, nor did we anything but doing
> reqular system upgrades using
> apt-get upgrade.
>
> HOwever, when we try a frequency breakdown, the following error
> appears (frst encountered about a week or so ago):
>
> CRITICAL ERROR - DATABASE CACHE OVERLOAD!
>
>
>
>
>
> CQPweb tried to clear database cache space but failed!
>
>
>
>
>
> Please report this error to the system administrator.
>
>
> PHP debugging backtrace
>
> array(5) {
>    [1]=>
>    array(4) {
>      ["file"]=>
>      string(35) "/var/www/html/cqpweb/lib/db.inc.php"
>      ["line"]=>
>      int(581)
>      ["function"]=>
>      string(20) "exiterror_dboverload"
>      ["args"]=>
>      array(0) {
>      }
>    }
>    [2]=>
>    array(4) {
>      ["file"]=>
>      string(35) "/var/www/html/cqpweb/lib/db.inc.php"
>      ["line"]=>
>      int(95)
>      ["function"]=>
>      string(16) "delete_saved_dbs"
>      ["args"]=>
>      array(0) {
>      }
>    }
>    [3]=>
>    array(4) {
>      ["file"]=>
>      string(42) "/var/www/html/cqpweb/lib/breakdown.inc.php"
>      ["line"]=>
>      int(160)
>      ["function"]=>
>      string(9) "create_db"
>      ["args"]=>
>      array(6) {
>        [0]=>
>        &string(4) "sort"
>        [1]=>
>        &string(10) "fqa49mvvxx"
>        [2]=>
>        &string(12) "[word="des"]"
>        [3]=>
>        &string(0) ""
>        [4]=>
>        &string(0) ""
>        [5]=>
>        &string(0) ""
>      }
>    }
>    [4]=>
>    array(4) {
>      ["file"]=>
>      string(41) "/var/www/html/cqpweb/lib/redirect.inc.php"
>      ["line"]=>
>      int(73)
>      ["args"]=>
>      array(1) {
>        [0]=>
>        string(42) "/var/www/html/cqpweb/lib/breakdown.inc.php"
>      }
>      ["function"]=>
>      string(7) "require"
>    }
>    [5]=>
>    array(4) {
>      ["file"]=>
>      string(37) "/var/www/html/cqpweb/exe/redirect.php"
>      ["line"]=>
>      int(1)
>      ["args"]=>
>      array(1) {
>        [0]=>
>        string(41) "/var/www/html/cqpweb/lib/redirect.inc.php"
>      }
>      ["function"]=>
>      string(7) "require"
>    }
> }
>
> and after a few more tries we finally reach the following error:
>
> Too many database processes!
>
> There are already 5 query sort databases being compiled.
>
> Please use the Back-button of your browser and try again in a few moments.
>
> Rebooting the server removes the latter error, but not the cache problem.
>
> What can be the cause of the cache problem and how can I remove it?
>
> Greetings from Saarbrücken,
>
> Jörg Knappen
>
>
> _______________________________________________
> CWB mailing list
> CWB at sslmit.unibo.it
> http://liste.sslmit.unibo.it/mailman/listinfo/cwb
> _______________________________________________
> CWB mailing list
> CWB at sslmit.unibo.it
> http://liste.sslmit.unibo.it/mailman/listinfo/cwb



_______________________________________________
CWB mailing list
CWB at sslmit.unibo.it
http://liste.sslmit.unibo.it/mailman/listinfo/cwb