We noticed that time to time that our Multiple journal, about 60 journals, OJS journal system is strained and causing the system to be not reachable. For instance yesterday, a Monday. Ours system people said the heaviest traffic came from LOCKSS. There is about 40 of the Journals that have the PKP PN plugin enabled. Some of those journals have a status of complete while others have in progress state for different issues. Is there anything that can be done? Is the only option to unable the plugin on some of the journals? Has anyone else had the problem?
Steps I took leading up to the issue
Not sure what to try.
What application are you using?
We are running 3.3.0.19.
Additional information
Please add any screenshots, logs or other information we can use to investigate this problem.
The completed deposits shouldn’t cause problems, for the others:
Pending deposits will issue a HTTP request to look for the status
New deposits are heavy, a .zip file will be generated for each issue
The operations are sequential, there’s no parallelism.
The preservation process can and should be improved (e.g. generating a .zip file could potentially be skipped), but it’s complex for us to define limits, this should be done on your side, you can setup CPU/disk/memory limits for the HTTP server/database.
Thanks @jonasraoni! Would putting limits on the CPU/disk/memory cause problems for uploading files or for the over all functionality of the OJS system? Thanks again.
It should run smoothly if the limits are setup properly, the hardware resources should be shared according to the system load (e.g. if you’ve got 10 users trying to load a page, a single process, like the PKP PN plugin creating deposits, shouldn’t be allowed to consume all of the disk bandwidth).
This is just a generic advice… Regarding the plugin, even if you don’t change anything, you shouldn’t have much problems with it, once the backlogged deposits are done, the server will get back to normal.