Bug#918751: Investigation results
severity 918751 minor retitle 918751 python-shade: Improvement of error message when failing to build on small machines (?) thanks Sorry, this was my fault. Here is the reason why it failed for me: This package (version 1.7.0) used to need 200 MB of RAM to build, and I was using a small machine (1 GB of RAM) to build it. So far, so good. The problem was that the current version (1.30.0) required around 1900 MB of RAM to build (a lot more than before) while I was still using small machines to build it. (You might be wondering why I'm using machines with 1 GB to build Debian packages: well, around 96.6% of all Debian packages in buster require 1 GB of RAM or less to build, and machines with 2 GB of RAM usually cost double the price per-hour). So no FTBFS problem here. Special thanks to Stewart Ferguson for actually looking at the build logs. Now I wonder if the error message in cases like this could be improved so that it's even more clear that lack of RAM was the problem. Probably it does not worth the effort, but I leave this to you. Thanks a lot.
Bug#918751: Investigation results
severity 827832 minor retitle 827832 python-shade: Improvement of error message when failing to build on small machines (?) thanks Sorry, this was my fault. Here is the reason why it failed for me: This package (version 1.7.0) used to need 200 MB of RAM to build, and I was using a small machine (1 GB of RAM) to build it. So far, so good. The problem was that the current version (1.30.0) required around 1900 MB of RAM to build (a lot more than before) while I was still using small machines to build it. (You might be wondering why I'm using machines with 1 GB to build Debian packages: well, around 96.6% of all Debian packages in buster require 1 GB of RAM or less to build, and machines with 2 GB of RAM usually cost double the price per-hour). So no FTBFS problem here. Special thanks to Stewart Ferguson for actually looking at the build logs. Now I wonder if the error message in cases like this could be improved so that it's even more clear that lack of RAM was the problem. Probably it does not worth the effort, but I leave this to you. Thanks a lot.
Bug#918751: Investigation results
owner 918751 sanv...@debian.org thanks I believe this problem is not directly related with the number of CPUs. Will do some additional tests and will probably downgrade this myself if I confirm my current suspicious. Thanks.
Bug#918751: Investigation results
noowner 827832 thank you The interesting part of the reporter's logs suggest that the following test fails: shade.tests.unit.test_object.TestObjectUploads.test_object_segment_retry_failure The failure is caused by a thread failing to spawn due to resource limitations on the reporter's machine. The reporter is using a single-processor VM. I tried to emulate this by using `dpkg-buildpackage -j1` and again by using `taskset --cpu-list 0 dpkg-buildpackage`. I was unable to reproduce the problem. I also fixed another RC bug in this package and buildd was also able to build this package. Because buildd and other single-core builds are all successful, I would regard this bug as `important` instead of `serious`. I am offloading ownership as I focus on other RC bugs.