<div dir="ltr"><div dir="ltr"><div class="gmail_default" style="font-size:small"><br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Tue, Feb 11, 2020 at 8:10 PM Jan Stancek <<a href="mailto:jstancek@redhat.com">jstancek@redhat.com</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><br>
<br>
----- Original Message -----<br>
> On Tue, Feb 11, 2020 at 6:52 PM Jan Stancek <<a href="mailto:jstancek@redhat.com" target="_blank">jstancek@redhat.com</a>> wrote:<br>
> <br>
> ><br>
> ><br>
> > ----- Original Message -----<br>
> > > On Mon, Feb 10, 2020 at 8:47 PM Jan Stancek <<a href="mailto:jstancek@redhat.com" target="_blank">jstancek@redhat.com</a>> wrote:<br>
> > ><br>
> > > > Lower the parameters so that test completes faster where possible.<br>
> > > ><br>
> > > > This also increases alarm timer slightly, which in combination with<br>
> > > > lower RLIMIT_CPU aims to avoid false positives in environments with<br>
> > > > high steal time, where it can take multiple of wall clock seconds<br>
> > > > to spend single second on a cpu.<br>
> > > ><br>
> > ><br>
> > > This patch could reduce the test failure possibility, but I'm afraid it<br>
> > > can't fix the problem radically, because with `stress -c 20' to overload<br>
> > an<br>
> > > s390x system(2cpus) in the background then setrlimit06(patched) still<br>
> > > easily gets failed:<br>
> > > setrlimit06.c:98: FAIL: Got only SIGXCPU after reaching both limit<br>
> > ><br>
> > > Another way I can think of is to raise the priority before its running,<br>
> > not<br>
> > > sure if that will disturb the original test but from my test, it always<br>
> > > gets a pass even with too much overload.<br>
> ><br>
> > Is this in addition to my patch? Because on its own I don't see how this<br>
> > will help when load is coming from different guests.<br>
> ><br>
> <br>
> Yes, this is only solving for itself loads. Besides the high steal time,<br>
> that's another reason I guess it causes the same failure, so do you think<br>
> it makes sense to merge two methods together?<br>
<br>
For now I'd go with just original patch. Until there is parallel test execution,<br>
there shouldn't be any local load during this test.<br></blockquote><div><br></div><div class="gmail_default" style="font-size:small">Ok sure. Let's apply the original first, then keep watching the status in the next testing.</div></div><div><br></div>-- <br><div dir="ltr" class="gmail_signature"><div dir="ltr"><div>Regards,<br></div><div>Li Wang<br></div></div></div></div>