[cfarm-users] GCC111 filesystem is very slow

[CM Graff]

Hey Everyone,

On a sort of tangential note, I often run my CI tests at -j1. The
problem with this is that it doesn't test parallel jobs, and I have
seen builds that succeed on -j1 that otherwise fail. Doesn't matter
much to me, but thought I'd throw in my two cents. I did see that
https://gitlab.com/git-vcs/git-gitlab-ci/blob/master/ci/gitlab/run-on-gcc-farm.sh
was reduced, so that is nice. There is also ulimit, but the
availability across machines is not generally very consistent. And
being that it's not POSIX specified as far as I know -- I recall
seeing some ulimit implementations with varying available options.


Graff

May there be peace and love for all of the peoples of the sacred
earth, such that we can explore mathematics and science as one.






On 11/25/18, Ævar Arnfjörð Bjarmason via cfarm-users
<cfarm-users at lists.tetaneutral.net> wrote:
> On Mon, Nov 26, 2018 at 1:46 AM Jeffrey Walton via cfarm-users
> <cfarm-users at lists.tetaneutral.net> wrote:
>> The filesystem on GCC111 seems very slow. Unpacking a tarball takes
>> minutes, and deleting the directory takes minutes. The tarball is
>> CMake sources, and it is only about 8 MB.
>>
>> Compiles of source files in the directory are relatively fast, so it
>> does not seem to be a compute bound problem.
>
> Hi. I believe this is my fault, sorry. Per
> <CAGWvny=_XXS1fO6f+Kd4WWhKd9K2TXtb==7rkpFu3AXr8eM2wg at mail.gmail.com>
> on this list ~1 hour ago I won't be doing git testing on gcc111, but
> hadn't killed the existing CI job I had there. Sorry again.
> _______________________________________________
> cfarm-users mailing list
> cfarm-users at lists.tetaneutral.net
> https://lists.tetaneutral.net/listinfo/cfarm-users
>