Oops didn't finish that email. If you have near concurrent publish operations 
under heavy traffic you may see this with slow modified_subnode updates. 

Regards, 
Andrew 



From: "Andrew Duck" <[email protected]> 
To: "Bertrand Dunogier" <[email protected]> 
Cc: "Stefan Georg" <[email protected]>, "sdk-public" 
<[email protected]> 
Sent: Thursday, 22 July, 2010 2:46:57 PM 
Subject: Re: [Sdk-public] DB lock failing to be set 


The query is slow - using those OR statements reads every row, it should be 
using an IN statement. Debug the query. 

Regards, 
Andrew 


andrew duck 
executive director | asia-pacific | quiqcorp.com 
155 high street | po box 6363 | christchurch 8011 new zealand 
p +64 3 341 7692 | f +64 3 341 7693 | m +64 212 934 985 (nz) | e 
[email protected] 


From: "Bertrand Dunogier" <[email protected]> 
To: "Stefan Georg" <[email protected]> 
Cc: "sdk-public" <[email protected]> 
Sent: Thursday, 22 July, 2010 2:44:42 PM 
Subject: Re: [Sdk-public] DB lock failing to be set 




2010/7/22 Stefan Georg < [email protected] > 


We are getting a few errors like this: 

Query error: Lock wait timeout exceeded; try restarting transaction. Query: 
UPDATE ezcontentobject_tree SET modified_subnode=1279768018 WHERE node_id='1' 
OR node_id='2' OR node_id='220' OR node_id='84458' 

once in a while, does it have to do with high DB traffic preventing the lock 
from being set. 

And if so, can the lock timeout be increased somehow? 

Increasing the lock timeout probably isn't the right solution. Check the 
following: 
- if using ezfind, are you using DelayedIndexing ? 
- has your SQL server been tuned correctly (innodb_buffer_pool_size, etc) ? 


-- 
Bertrand 

-- 
Sdk-public mailing list 
[email protected] 
http://lists.ez.no/mailman/listinfo/sdk-public 

-- 
Sdk-public mailing list 
[email protected] 
http://lists.ez.no/mailman/listinfo/sdk-public 
-- 
Sdk-public mailing list
[email protected]
http://lists.ez.no/mailman/listinfo/sdk-public

Reply via email to