Re: Centos : Load Average :OS Process Name : migration/1, migration/2 , migration/n

On Thu, May 23, 2019 at 9:18 AM Perumal Raj <perucinci(at)gmail(dot)com> wrote:
> We have recently migrated postgres DB to out of of server ( Centos 6.9 ) .
> Both Source and Target versions of OS/DB are same . Also Configuration is Apple-Apple.
>
> But We started seeing lot of process name 'migration' at OS Level in new server which is triggering Load average most of the time .
>
> Some sample process from top command,
>
> 67 root RT 0 0 0 0 R 34.8 0.0 913:06.38 [migration/16]
> 155 root RT 0 0 0 0 S 32.2 0.0 845:46.53 [migration/38]
> 35 root RT 0 0 0 0 S 27.6 0.0 927:15.27 [migration/8]
> 11 root RT 0 0 0 0 S 21.4 0.0 1033:45 [migration/2]
> 131 root RT 0 0 0 0 S 21.4 0.0 812:00.70 [migration/32]
> 87 root RT 0 0 0 0 S 20.7 0.0 394:28.32 [migration/21]
>
> Anyone come-across this situation ? Please share your thoughts .

I don't know anything about this, but I found a claim (without much
real explanation) that Linux < 3.6.11 had some kind of problem in this
area, could be relevant:

https://serverfault.com/questions/674685/kernel-processes-periodically-eating-cpu-during-high-load
https://bugs.gentoo.org/394487
https://bugzilla.kernel.org/show_bug.cgi?id=47341

--
Thomas Munro
https://enterprisedb.com