[Users] BUG: soft lockup

[suporte at logicworks.pt]

Well, this happens a second time. I did an upgrade to the system, engine and fedora18 hosts, oVirt Engine Version: 3.2.1-1.fc18. Now I have a 2 Fujitsu servers. Both with kernel 3.8.2 - 206.fc18.x86_64, CPU one host : Intel(R) Xeon(R) CPU X3430  @ 2.40GHz (SMT Disabled), CPU other host : Intel(R) Xeon(R) CPU X3450 (SMT Enabled), installed packages on both hosts:
vdsm-python-4.10.3-9.fc18.x86_64
vdsm-gluster-4.10.3-9.fc18.noarch
vdsm-cli-4.10.3-9.fc18.noarch
vdsm-xmlrpc-4.10.3-9.fc18.noarch
vdsm-4.10.3-9.fc18.x86_64

When I add the second host start having this errors, BUG: soft lockup, and the host died.

Should a kernel downgrade solve the problem?



----- Mensagem original -----
De: "Gianluca Cecchi" <gianluca.cecchi at gmail.com>
Para: suporte at logicworks.pt
Cc: "users" <Users at ovirt.org>
Enviadas: Sexta-feira, 15 Março, 2013 18:42:12
Assunto: Re: [Users] BUG: soft lockup

On Fri, Mar 15, 2013 at 7:21 PM,  wrote:
> I have a host that have this error BUG: soft lockup - CPU#4 stuck for 22s!
> [sh: 1534], and the host died
>
> this host was installed from fedora18 minimum installation, kernel version
> 3.8.2-206.fc18.x86_64. I have another host with the same configuration, and
> have no errors, until now.
>
> Any idea?

I can only say that I hadd the same problem some days ago when I tried
to add a new node to a pre-existing oVirt 3.2 install.
I erroneously supposed it was a problem with my hw but when I tried to
add another same hw node it got the same error.
It was your same kernel giving problems.
I solved the problem installing the same kernel as the first node,
that was the latest 3.7.x in F18 release.
Possibly 3.8.2 having problems with some cpus.
My nodes were HP blades BL685c G1 with AMD Opteron G2 processors.

Not had time to open bug or query on F18 user list yet...
Yours?

I also have a running oVirt 3.2 all-in-one setup with Fedora 18 and
kernel 3.8.1-201.fc18.x86_64.
In this case the cpu is (from cpuinfo) AMD Athlon(tm) II X4 630
Processor and configured in oVirt as "AMD Opteron G3"


I'm going to upgrade this install with these steps:
1) only kernel and see what happens
2) upgrade to 3.2.1

Gianluca