cn at brain-biotech.de
2021-Feb-28 15:37 UTC
[Samba] Time outs and high cpu usage after upgrading to 4.12.11
Hello we have seen this also on 4.12 and Centos8. Only the upgrade to 4.13 fixed the problem for us. Regards Christian -- Dr. Christian Naumer Vice President Unit Head Bioprocess Development B.R.A.I.N Aktiengesellschaft Darmstaedter Str. 34-36, D-64673 Zwingenberg e-mail cn at brain-biotech.com, homepage www.brain-biotech.com phone +49-6251-9331-30 / fax +49-6251-9331-11 Sitz der Gesellschaft: Zwingenberg/Bergstrasse Registergericht AG Darmstadt, HRB 24758 Vorstand: Adriaan Moelker (Vorstandsvorsitzender), Lukas Linnig Aufsichtsratsvorsitzender: Dr. Georg Kellinghusen
Andrew Bartlett
2021-Feb-28 21:42 UTC
[Samba] Time outs and high cpu usage after upgrading to 4.12.11
On Sun, 2021-02-28 at 16:37 +0100, cn--- via samba wrote:> Hello we have seen this also on 4.12 and Centos8. Only the upgrade > to > 4.13 fixed the problem for us.That's really interesting. I'm not aware of a deliberate change or fix, I would be very keen to know if a particular commit that addresses this is ever identified, so we can learn and perhaps lock in the correct behaviour with a test. Not that I'm suggesting such tests in production, but if anybody here can reproduce this well enough and has the time to do a git bisect, either from the old working to the first failing version or from the first failing version to the new working one that would be awesome. Andrew Bartlett -- Andrew Bartlett (he/him) https://samba.org/~abartlet/ Samba Team Member (since 2001) https://samba.org Samba Team Lead, Catalyst IT https://catalyst.net.nz/services/samba Samba Development and Support, Catalyst IT - Expert Open Source Solutions
cn at brain-biotech.de
2021-Mar-01 08:44 UTC
[Samba] Time outs and high cpu usage after upgrading to 4.12.11
Hello all, as we upgraded all our DCs to 4.13 before I could gather all the data for a bug report, here is what we have seen. -- We are using the Sernet Packages -- 4.12 on Centos 7 did not show this behaviour -- The problem did not show after a reboot. It took 6-7 days running and the CPU load started to increase then. I saw some time outs on slower VMs. There I increased the vCPU number to 8 and I could prolong the time until there was a noticeable impact (Login by ssh took very long). -- After a restart of the service everything went back to normal for 6-7 days). Then again CPU load increased continuously. -- Disabling audit logging also prolonged the time until we saw any impact. But that probably just reduced the load. -- I suspected it has something todo with gnu-tls but could never prove it. Regards Christian Am 28.02.21 um 22:42 schrieb Andrew Bartlett via samba:> On Sun, 2021-02-28 at 16:37 +0100, cn--- via samba wrote: >> Hello we have seen this also on 4.12 and Centos8. Only the upgrade >> to >> 4.13 fixed the problem for us. > > That's really interesting. I'm not aware of a deliberate change or > fix, I would be very keen to know if a particular commit that addresses > this is ever identified, so we can learn and perhaps lock in the > correct behaviour with a test. > > Not that I'm suggesting such tests in production, but if anybody here > can reproduce this well enough and has the time to do a git bisect, > either from the old working to the first failing version or from the > first failing version to the new working one that would be awesome. > > Andrew Bartlett >-- Dr. Christian Naumer Vice President Unit Head Bioprocess Development B.R.A.I.N Aktiengesellschaft Darmstaedter Str. 34-36, D-64673 Zwingenberg e-mail cn at brain-biotech.com, homepage www.brain-biotech.com phone +49-6251-9331-30 / fax +49-6251-9331-11 Sitz der Gesellschaft: Zwingenberg/Bergstrasse Registergericht AG Darmstadt, HRB 24758 Vorstand: Adriaan Moelker (Vorstandsvorsitzender), Lukas Linnig Aufsichtsratsvorsitzender: Dr. Georg Kellinghusen