Are you responsible for making OJS work -- installing, upgrading, migrating or troubleshooting? Do you think you've found a bug? Post in this forum.
Moderators: jmacgreg, btbell, michael, bdgregg, barbarah, asmecher
The Public Knowledge Project Support Forum is moving to http://forum.pkp.sfu.ca
This forum will be maintained permanently as an archived historical resource, but all new questions should be added to the new forum. Questions will no longer be monitored on this old forum after March 30, 2015.
Is there anything we can do to improve performance when searching over all 250 journals? We loaded +/- 40 000 articles. I've checked that indexes are build in MYSQL. Our installation of OJS 2.1.1 is running on a linux server:
Linux zypher 2.6.17-11-server
processor : 7
cpu MHz : 3192.109
Response time on searching can be tested on:
- Posts: 16
- Joined: Thu Sep 07, 2006 2:26 am
- Location: Pretoria
Searching is going to be an IO-limited operation, so make sure you focus your attention on that. From a couple of test searches, the performance looks acceptable to me; you may find that you're able to alter a few of the foreign key and primary key data types in the database to be smaller, which will also compact the indexes. If you have any database optimization experience, you may want to experiment with different storage engines and in-memory strategies.
At a certain number of records, a reverse MySQL index like the one used in OJS will start to degrade. We intend to assess Lucene in the future for possible integration with the Harvester application; if it turns out to be a straight-forward process we may offer it as an option in OCS and OJS. However, the current approach has performed fine for those applications in our experience thus far.
- Posts: 10015
- Joined: Wed Aug 10, 2005 12:56 pm
Return to OJS Technical Support
Who is online
Users browsing this forum: Yahoo [Bot] and 2 guests