pnpm killed
Any ideas on what this could mean? Literally only happens in the build container somehow.
Why and how does the process get killed?
15 Replies
Oh scratch that, it does:
Its weird, it just happens sometimes
Could it be due to memory not being allocated properly? When i run it multiple times the available memory goes up, and after a while the install succeeds
@Michal Saloň @Matěj Pavlíček
Hi, if you can run it locally, how much RAM does it use? Because our build containers have fixed 8GB of RAM, so if it needs more, it will fail.
You can test this using the
Pipeline debug options with Before first command, just like you did, but connect to the build container twice.
Run top or htop (to add htop you need to run sudo apk add htop first) on one tab, start your commands in second tab, and look at the RAM usage in top/htop (you can sort by RAM).
If it somehow goes up to 8GB, then that would be the issue. If you see that the container has way less RAM than 8GB as it should, that is a bug on our side.
If more RAM is needed, I can adjust it for that one build container to see if it helps, but 8GB should be plenty for 5CPU cores.I can build it under my VM which has 4GBs
I am testing it and can't reproduce it. Container always starts with 8GB of RAM, whether it is new, or existing one with cache (see the
available column):
The container it looks like starts with ~300MB
This is on debug before the first command
Somehow it went down
its really sporadic
Also it seems like im not the only one: https://discord.com/channels/735781031147208777/735781031663239190/1344095677977727000, pinging @Char
IIRC Char had issue with a missing package?
Leave the container running for a bit, le tme check what it's set to
No, it was also killed but in a different stage e.g. build and not package install
Should be fixed :skillIssue:
Looks like its fne, thanks!