** Also affects: stress-ng
Importance: Undecided
Status: New
** Changed in: stress-ng
Importance: Undecided => High
** Changed in: stress-ng
Status: New => Fix Committed
** Changed in: stress-ng
Assignee: (unassigned) => Colin Ian King (colin-king)
** Changed in: linux
Added some more smarts to make this a non-failure issue, just a resource
limiting exit on each stressor:
http://kernel.ubuntu.com/git/cking/stress-
ng.git/commit/?id=d3db4119cdd3bee439a4f2a6cc36d27ff3ce6c4b
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is s
NOTE: I'll fix up stress-ng to detect this and report it is a problem,
but I do think the failure is correct - stress-ng hit an upper resource
bound set by the system and failed. However, it should be reporting that
the errno meant that there were no resources left.
--
You received this bug notif
Ah, looks like for large multiprocessor machines, /proc/sys/fs/aio-max-
nr is too low and causing the setup to fail because of the low default
for so many concurrent stressors.
I recommend writing 1000 * number of CPUs into /proc/sys/fs/aio-max-nr
before running the aiol test. That should give it
** Also affects: plainbox-provider-checkbox (Ubuntu)
Importance: Undecided
Status: New
** No longer affects: linux (Ubuntu)
** Changed in: plainbox-provider-checkbox (Ubuntu)
Status: New => Confirmed
--
You received this bug notification because you are a member of Ubuntu
Bugs,
** Summary changed:
- Disk test fails on IBM Power S822LC (8335-GTB)
+ stress-ng tests failing
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1640547
Title:
stress-ng tests failing
To manage notifi
OK, I can see one big problem, the timeout command is being used to
SIGKILL stress-ng - that means all the child processes get SIGKILL'd and
hence abort without cleaning up. One ends up with a file system full of
garbage files which will cause subsequent tests to break.
I suggest:
1. Sending SIG
