We need to replace the volume hosting our RPM/YUM repo data with a larger one.
As a consequence, we'll have to stop the build queue processing for about 20
hours, and we'll have two "full" (but short) outages (2x5 minutes) when the
yum
repositories will be unavailable.
The anticipated start of the outage (queue stops processing) is Saturday 2022-10-29:
$ date --date "2022-10-29 17:00 UTC"
Sat Oct 29 07:00:00 PM CEST 2022
But a few hours before, we'll lower the build throughput a bit to prioritize the
I/O we have on the data migration task..
We'd like to kindly ask you to minimize the build throughput in the Fedora
Copr build service during the weekend in question — just to allow us to
make the transition as smooth as possible. It would be ideal if you could
e.g. delay your mass rebuilding and other large data modifications tasks
until the storage movement is over. That said, we don't plan to turn the
frontend machine off (API and web-UI) during the whole outage period.
For more info, see the more detailed schedule:
https://docs.google.com/spreadsheets/d/1URopuw2R533H6i4vSTH8Nt47EIhs3ubWc...
Affected services:
- Copr service (frontend):
https://copr.fedorainfracloud.org
- Copr (backend) repositories:
https://download.copr.fedorainfracloud.org/ (AWS CloudFront proxy)
https://copr-be.cloud.fedoraproject.org/ (storage)
Sorry for the inconvenience,
Copr Team