On our OOD 3.0 instance I am seeing this error in the logs:
App 2944987 output: [2023-07-05 14:34:22 -0600 ] ERROR "Error Errno::ENOENT when reading and parsing quota file https://www.chpc.utah.edu/apps/systems/curl_post/quota.json for user u0101881: No such file or directory @ rb_sysopen - https://www.chpc.utah.edu/apps/systems/curl_post/quota.json"
I am wondering if there’s an issue with that in 3.0, since in 2.0, I get this warning, but the quota are working:
App 130518 output: /var/www/ood/apps/sys/dashboard/app/models/quota.rb:20: warning: calling URI.open via Kernel#open is deprecated, call URI.open directly or use URI#open
Looking at this and thinking about the change from ruby 2.7 to ruby 3 the deprecation warning that is calling out a change with open-uri being used there which is an extension of Kernel#open. It shouldn’t be effecting the functionality of anything though.
I’ve opened an issue to fix the deprecation warning though, thanks for pointing that out!
thanks for looking at this. I am more concerned about the first error message that comes with 3.0 which complains that it can’t open the http json file - it sounds like the new open ruby function does not do that on remote files over http.
I just set the OOD_QUOTA_THRESHOLD=0.1, which usually shows some warnings for me but in our 3.0 test instance it does not show anything.
Thanks Martin, I’ll have to take a look more as I’m not familiar with this piece of the software or what is going on here, but thanks for bringing this up and I’ll try to see what I can find.
Hi Travis, after doing some more comparisons with our 2.0 production systems, I can definitely say that the 3.0 http download of the quota json file does not work, as in 2.0 the quotas get displayed (though with the warning I sent earlier), while in 3.0 they don’t (and we get the error).
This is quite an useful feature since it gives users an advance warning when their quota is close to be full, so, I appreciate if you find time soon to check this.
Update on this, it’s not critical anymore for us to fix the http pull since we ended up setting a cron job to get the json files locally, after realizing that they get pulled over http every time someone logs in, while the files change less often. The local json file read works fine.