Describe the issue or problem
Hello everyone!
We are facing slowness when using sidebar filters of dashboard.
The database has more than 10.000 submissions/publications.
I’m wondering about installing APC cache. What do you think?
Does anyone have any suggestion?
Steps I took leading up to the issue
In the dashboard, in active submissions tab, use the filter.
What application are you using?
OJS 3.4.0-1
Hi @liviagouvea,
APC has been discontinued, I’d use Zend OPcache, it gives a good bump on PHP performance.
Best regards,
Josh
What about Memcache?
Hi all,
OJS doesn’t currently use Memcache effectively (and in fact we temporarily removed that support from recent releases pending a proper reassessment).
@liviagouvea, it is a good idea to use an opcache like Zend Opcache, but you might find that OJS 3.4.0 still performs poorly on large submission lists. I’d suggest testing with 3.5.0; we’ve rewritten the submission lists entirely, including testing on larger datasets, and any further improvements we make will be based on that line of releases. I suspect you’ll already be much better off, not to mention that the presentation is much improved.
If your DBMS is on a different server than the application server, you might consider moving them to the same machine. OJS is known to execute a lot of database queries for certain pages, and the overhead of connecting over the network can add up.
Regards,
Alec Smecher
Public Knowledge Project Team
Thank you, @asmecher
In fact, I think IA scraping bots are also affecting the performance.
Hi @liviagouvea,
If you’ve enabled the Recommend Similar Articles plugin, that’s known to come with heavy indexing load; see:
https://github.com/pkp/pkp-lib/issues/9897
If you’re using that plugin, I’d recommend disabling it to see if that helps. Otherwise, reviewing your access logs to see what kind of bot behaviour you’re encountering is probably a good idea; you can use a mixture of robots.txt (for well-behaved bots) and IP-based blocking (for less well-behaved ones) to tune your exposure.
Thanks,
Alec Smecher
Public Knowledge Project Team