CentOS - May 2019 - is "list_del corruption" fix available in Centos ?

Hi,

Gentle reminder ! Please let me know if there are any pointers for this

Thanks
Santhosh

On Tue, May 14, 2019 at 7:26 PM santhosh kumar <santhosh.santuu at
gmail.com>
wrote:
> Hi,
>
> We migrated from redhat 5.3 to centos 7.5 and facing crashes in longevity
> tests
>
> All of them point to below reason,
>
> list_del corruption. next->prev should be ffff880c1e567000, but was
> 00450008a948adba
>
> We searched around web and see this is fixed in redhat
> https://bugzilla.redhat.com/show_bug.cgi?id=1028750
>
> But don't see any fix in Centos.
https://bugs.centos.org/view.php?id=10944
>
> Is this the fact ?
> What is the general practice ?
> Can we port the fix from redhat and upstream for Centos ?
>
> Please give me some guidance.
>
> Thanks
> Santhosh
>

> 
> Gentle reminder ! Please let me know if there are any pointers for this
As far as I can see your original message never made it on to the
mailing list ...
> 
> On Tue, May 14, 2019 at 7:26 PM santhosh kumar <santhosh.santuu at
gmail.com>
> wrote:
> 
> > Hi,
> > 
> > We migrated from redhat 5.3 to centos 7.5 and facing crashes in
longevity
> > tests
Why not 7.6??
> > 
> > All of them point to below reason,
> > 
> > list_del corruption. next->prev should be ffff880c1e567000, but was
> > 00450008a948adba
> > 
> > We searched around web and see this is fixed in redhat
> > https://bugzilla.redhat.com/show_bug.cgi?id=1028750
> > 
> > But don't see any fix in Centos.
https://bugs.centos.org/view.php?id=10944
> > 
> > Is this the fact ?
> > What is the general practice ?
> > Can we port the fix from redhat and upstream for Centos ?
> > 
CentOS is a clone of RHEL - if it is fixed in RHEL 7 it will be fixed
in CentOS 7. CentOS doesn't "fix" things as such as that would
break
compatibility with RHEL.

There may be some 3rd party repo that provides a newer kernel that
fixes the issue.

P.

On Fri, 17 May 2019, Pete Biggs wrote:
> CentOS is a clone of RHEL - if it is fixed in RHEL 7 it will be fixed
> in CentOS 7. CentOS doesn't "fix" things as such as that
would break
> compatibility with RHEL.
>
> There may be some 3rd party repo that provides a newer kernel that
> fixes the issue.
I'd go further.

It's also a kernel bug in btrfs, which given btrfs has only ever been a
Technology Preview in RHEL, and was moved to being a deprecated feature in 7.4,
I'd suggest that you shouldn't be looking for upstream support on btrfs.

If you're tinkering with btrfs, don't be surprised at having to use a
kernel from elsewhere, as suggested by Pete.

RHEL advice would clearly be not to use btrfs.

jh

On Fri, May 17, 2019 at 2:47 AM santhosh kumar
<santhosh.santuu at gmail.com> wrote:
> On Tue, May 14, 2019 at 7:26 PM santhosh kumar <santhosh.santuu at
gmail.com>
> wrote:
>
> > We migrated from redhat 5.3 to centos 7.5 and facing crashes in
longevity
> > tests
> >
> > All of them point to below reason,
> >
> > list_del corruption. next->prev should be ffff880c1e567000, but was
> > 00450008a948adba
> >
> > We searched around web and see this is fixed in redhat
> > https://bugzilla.redhat.com/show_bug.cgi?id=1028750
That's a Fedora bug and Fedora merely built a newer kernel with the
upstream kernel fix.  It's mostly irrelevant to RHEL and Centos.
> >
> > But don't see any fix in Centos.
https://bugs.centos.org/view.php?id=10944
According to the CentOS kernel changelog

$ rpm -q --changelog kernel-`uname -r`|less # only relevant lines shown
* Mon Mar 03 2014 Jarod Wilson <jarod at redhat.com> [3.10.0-101.el7]
- [fs] btrfs: take ordered root lock when removing ordered operations
inode (Zach Brown) [1051282]

this might have been fixed in 2014, but I don't have access to
1051282.  There are also scores more btrfs patches in the RHEL 7
kernel since 2013.

I also notice that the stack trace for
https://bugs.centos.org/view.php?id=10944 does NOT mention btrfs, so
it's most likely a different bug.

Jim