-
Notifications
You must be signed in to change notification settings - Fork 9
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Data taring before downloads will be really slow #64
Comments
nikromen
added a commit
to nikromen/log-detective-website
that referenced
this issue
Jan 7, 2024
This solution raises two concerns for the future once we have a lot of data collected: - it will be a while until the data will be tarred - creating delay before actual download (after cca 200 feedbacks relly noticable delay :/) look at example with only 300 (tiny!!) feedbacks: [root@backend persistent]# time tar -zcf results.tar.gz results/ real m9.473s user m9.328s sys m0.483s - downloading takes also some time -> thus blocking the whole worker during this. IIRC we have 8 workers thus 8 downloads and API is unresponsive. Solution: - how to solve the delay before download: fedora-copr#64 - the issue above is the slowest, once that will be resolved and people still complain, do this: fedora-copr#65
nikromen
changed the title
Data downloads will be really slow
Data taring before downloads will be really slow
Jan 7, 2024
Merged
nikromen
added a commit
that referenced
this issue
Jan 8, 2024
This solution raises two concerns for the future once we have a lot of data collected: - it will be a while until the data will be tarred - creating delay before actual download (after cca 200 feedbacks relly noticable delay :/) look at example with only 300 (tiny!!) feedbacks: [root@backend persistent]# time tar -zcf results.tar.gz results/ real m9.473s user m9.328s sys m0.483s - downloading takes also some time -> thus blocking the whole worker during this. IIRC we have 8 workers thus 8 downloads and API is unresponsive. Solution: - how to solve the delay before download: #64 - the issue above is the slowest, once that will be resolved and people still complain, do this: #65
We should prioritize this once this really causes issues. |
nikromen
added a commit
to nikromen/log-detective-website
that referenced
this issue
Jul 29, 2024
This solution raises two concerns for the future once we have a lot of data collected: - it will be a while until the data will be tarred - creating delay before actual download (after cca 200 feedbacks relly noticable delay :/) look at example with only 300 (tiny!!) feedbacks: [root@backend persistent]# time tar -zcf results.tar.gz results/ real m9.473s user m9.328s sys m0.483s - downloading takes also some time -> thus blocking the whole worker during this. IIRC we have 8 workers thus 8 downloads and API is unresponsive. Solution: - how to solve the delay before download: fedora-copr#64 - the issue above is the slowest, once that will be resolved and people still complain, do this: fedora-copr#65
nikromen
added a commit
to nikromen/log-detective-website
that referenced
this issue
Jul 29, 2024
This solution raises two concerns for the future once we have a lot of data collected: - it will be a while until the data will be tarred - creating delay before actual download (after cca 200 feedbacks relly noticable delay :/) look at example with only 300 (tiny!!) feedbacks: [root@backend persistent]# time tar -zcf results.tar.gz results/ real m9.473s user m9.328s sys m0.483s - downloading takes also some time -> thus blocking the whole worker during this. IIRC we have 8 workers thus 8 downloads and API is unresponsive. Solution: - how to solve the delay before download: fedora-copr#64 - the issue above is the slowest, once that will be resolved and people still complain, do this: fedora-copr#65
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
We can't create cronjob periodically taring the result directory so the /download endpoint itself does this job. This is horribly slow, look at the example with only 300 really small feedbacks (just name, solution and reason for only 1 log):
and we want thousands of feedbacks...
This could be solved by handling the taring job to someone else - cronjob in openshift, but we need RWX PVC so containers can share the volume between them. We don't have permission to create this kind of volume in our openshift console so we have to ask someone. Reach out to someone who can permit us to create RWX PVC (fedora infra folks?).
The cronjob solution itself was already written so you can copy-paste it from nikromen@4381ec2
The text was updated successfully, but these errors were encountered: