Hi Guys
I'm trying to add the patch mentioned in this thread to the plain Centos kernel
http://lists.centos.org/pipermail/centos/2008-August/062845.html
I've followed the directions on the wiki for building a custom kernel, but after applying all the patches it fails with:
make[1]: *** [nonint_oldconfig] Error 4 make: *** [nonint_oldconfig] Error 2 error: Bad exit status from /var/tmp/rpm-tmp.74263 (%prep)
RPM build errors: Bad exit status from /var/tmp/rpm-tmp.74263 (%prep)
Can anyone help with this? Or is there an easier way to apply the patch, as I'm not wanting to change the config of the kernel in any way?
Thanks
Dunc
On Wed, Sep 3, 2008 at 3:37 AM, Dunc centos@duncb.co.uk wrote:
Hi Guys
I'm trying to add the patch mentioned in this thread to the plain Centos kernel
http://lists.centos.org/pipermail/centos/2008-August/062845.html
I've followed the directions on the wiki for building a custom kernel, but after applying all the patches it fails with:
make[1]: *** [nonint_oldconfig] Error 4 make: *** [nonint_oldconfig] Error 2 error: Bad exit status from /var/tmp/rpm-tmp.74263 (%prep)
RPM build errors: Bad exit status from /var/tmp/rpm-tmp.74263 (%prep)
Can anyone help with this? Or is there an easier way to apply the patch, as I'm not wanting to change the config of the kernel in any way?
Just wanted to make sure you ran the rpmbuild command with --target=`uname -m` ? Or did you attempt to build it for an arch that is different from the running kernel?
Now...
This bug looks quite nasty and seems to be affecting a good number of people. There is a long discussion thread on the Scientific Linux and someone there is offering patched SL kernels:
http://listserv.fnal.gov/scripts/wa.exe?A2=ind0808&L=scientific-linux-us...
In a case like this, Johnny Hughes used to build a patched version of CentOS kernels. Seems like he is not available at this moment. Would some other CentOS dev be able to do this for affected CentOS users?
Akemi
Akemi Yagi wrote:
On Wed, Sep 3, 2008 at 3:37 AM, Dunc centos@duncb.co.uk wrote:
Hi Guys
I'm trying to add the patch mentioned in this thread to the plain Centos kernel
http://lists.centos.org/pipermail/centos/2008-August/062845.html
I've followed the directions on the wiki for building a custom kernel, but after applying all the patches it fails with:
make[1]: *** [nonint_oldconfig] Error 4 make: *** [nonint_oldconfig] Error 2 error: Bad exit status from /var/tmp/rpm-tmp.74263 (%prep)
RPM build errors: Bad exit status from /var/tmp/rpm-tmp.74263 (%prep)
Can anyone help with this? Or is there an easier way to apply the patch, as I'm not wanting to change the config of the kernel in any way?
Just wanted to make sure you ran the rpmbuild command with --target=`uname -m` ? Or did you attempt to build it for an arch that is different from the running kernel?
Yes, i was following the wiki guide which uses that command
Now...
This bug looks quite nasty and seems to be affecting a good number of people. There is a long discussion thread on the Scientific Linux and someone there is offering patched SL kernels:
http://listserv.fnal.gov/scripts/wa.exe?A2=ind0808&L=scientific-linux-us...
In a case like this, Johnny Hughes used to build a patched version of CentOS kernels. Seems like he is not available at this moment. Would some other CentOS dev be able to do this for affected CentOS users?
That would be brilliant..
Akemi
cheers
Dunc
On Wed, Sep 03, 2008 at 04:56:06PM +0100, Dunc wrote:
Akemi Yagi wrote:
On Wed, Sep 3, 2008 at 3:37 AM, Dunc centos@duncb.co.uk wrote:
Hi Guys
I'm trying to add the patch mentioned in this thread to the plain Centos kernel
http://lists.centos.org/pipermail/centos/2008-August/062845.html
...
This bug looks quite nasty and seems to be affecting a good number of people. There is a long discussion thread on the Scientific Linux and someone there is offering patched SL kernels:
http://listserv.fnal.gov/scripts/wa.exe?A2=ind0808&L=scientific-linux-us...
In a case like this, Johnny Hughes used to build a patched version of CentOS kernels. Seems like he is not available at this moment. Would some other CentOS dev be able to do this for affected CentOS users?
That would be brilliant..
Here it goes (testing key signed and not gone through the regular built system): http://people.centos.org/tru/kernel+bz453094/
Cheers,
Tru
Tru Huynh wrote
Here it goes (testing key signed and not gone through the regular built system): http://people.centos.org/tru/kernel+bz453094/
Cheers,
Tru
Thank you, thats great. From what i can tell looking at the .spec it is almost identical to the one I have built (although I only built base). I've been running my kernel over night and hammering the nfs with no errors, so looks like this has solved the issue.
What is the best way of letting people know about the availability of this kernel, as it must be a quite widespread problem?
cheers
Dunc
On Thu, Sep 4, 2008 at 1:52 AM, Duncan centos@duncb.co.uk wrote:
Tru Huynh wrote
Here it goes (testing key signed and not gone through the regular built system): http://people.centos.org/tru/kernel+bz453094/ Cheers,
Tru
Thank you, thats great. From what i can tell looking at the .spec it is almost identical to the one I have built (although I only built base). I've been running my kernel over night and hammering the nfs with no errors, so looks like this has solved the issue.
What is the best way of letting people know about the availability of this kernel, as it must be a quite widespread problem?
According to the upstream bugzilla, the fix will be in kernel-2.6.18-92.1.13.el5. We don't know how soon it will come out but ...
Akemi
On Thu, Sep 4, 2008 at 7:16 AM, Akemi Yagi amyagi@gmail.com wrote:
On Thu, Sep 4, 2008 at 1:52 AM, Duncan centos@duncb.co.uk wrote:
Tru Huynh wrote
Here it goes (testing key signed and not gone through the regular built system): http://people.centos.org/tru/kernel+bz453094/
Tru
Thank you, thats great. From what i can tell looking at the .spec it is almost identical to the one I have built (although I only built base). I've been running my kernel over night and hammering the nfs with no errors, so looks like this has solved the issue.
According to the upstream bugzilla, the fix will be in kernel-2.6.18-92.1.13.el5. We don't know how soon it will come out but ...
A bit late followup but... the 2.6.18-92.1.13.el5 kernel is out.
Akemi
Dunc wrote:
Hi Guys
I'm trying to add the patch mentioned in this thread to the plain Centos kernel
http://lists.centos.org/pipermail/centos/2008-August/062845.html
I've followed the directions on the wiki for building a custom kernel, but after applying all the patches it fails with:
make[1]: *** [nonint_oldconfig] Error 4 make: *** [nonint_oldconfig] Error 2 error: Bad exit status from /var/tmp/rpm-tmp.74263 (%prep)
RPM build errors: Bad exit status from /var/tmp/rpm-tmp.74263 (%prep)
Just to follow up on this, the patch meant nonint_oldconfig was expecting some more config options in the .config.
I had to add:
CONFIG_PATA_ISAPNP=n CONFIG_PATA_LEGACY=n CONFIG_PATA_QDI=n CONFIG_PATA_WINBOND_VLB=n
to the base and PAE configs, and then the build continued. No idea if putting n was the right option, but I'm sure i'll find out when I run the kernel..
cheers
Dunc
On Wed, Sep 3, 2008 at 9:31 AM, Dunc centos@duncb.co.uk wrote:
Dunc wrote:
Hi Guys
I'm trying to add the patch mentioned in this thread to the plain Centos kernel
http://lists.centos.org/pipermail/centos/2008-August/062845.html
I've followed the directions on the wiki for building a custom kernel, but after applying all the patches it fails with:
make[1]: *** [nonint_oldconfig] Error 4 make: *** [nonint_oldconfig] Error 2 error: Bad exit status from /var/tmp/rpm-tmp.74263 (%prep)
RPM build errors: Bad exit status from /var/tmp/rpm-tmp.74263 (%prep)
Just to follow up on this, the patch meant nonint_oldconfig was expecting some more config options in the .config.
I had to add:
CONFIG_PATA_ISAPNP=n CONFIG_PATA_LEGACY=n CONFIG_PATA_QDI=n CONFIG_PATA_WINBOND_VLB=n
to the base and PAE configs, and then the build continued. No idea if putting n was the right option, but I'm sure i'll find out when I run the kernel..
Hmm, that is odd. If you copy the config file from configs/ directory, make nonint_oldconfig should run without any needing for edit.
Akemi
Akemi Yagi wrote:
On Wed, Sep 3, 2008 at 9:31 AM, Dunc centos@duncb.co.uk wrote:
Dunc wrote:
Hi Guys
I'm trying to add the patch mentioned in this thread to the plain Centos kernel
http://lists.centos.org/pipermail/centos/2008-August/062845.html
I've followed the directions on the wiki for building a custom kernel, but after applying all the patches it fails with:
make[1]: *** [nonint_oldconfig] Error 4 make: *** [nonint_oldconfig] Error 2 error: Bad exit status from /var/tmp/rpm-tmp.74263 (%prep)
RPM build errors: Bad exit status from /var/tmp/rpm-tmp.74263 (%prep)
Just to follow up on this, the patch meant nonint_oldconfig was expecting some more config options in the .config.
I had to add:
CONFIG_PATA_ISAPNP=n CONFIG_PATA_LEGACY=n CONFIG_PATA_QDI=n CONFIG_PATA_WINBOND_VLB=n
to the base and PAE configs, and then the build continued. No idea if putting n was the right option, but I'm sure i'll find out when I run the kernel..
Hmm, that is odd. If you copy the config file from configs/ directory, make nonint_oldconfig should run without any needing for edit.
You would think that, but the configs in the config directory do not contain those options, and it wouldn't build without them.
either way the build has just failed with:
"make: *** [drivers] Error 2 + exit 1"
any ideas?
Dunc
On Wed, Sep 3, 2008 at 10:27 AM, Dunc centos@duncb.co.uk wrote:
Akemi Yagi wrote:
Just to follow up on this, the patch meant nonint_oldconfig was expecting some more config options in the .config.
I had to add:
CONFIG_PATA_ISAPNP=n CONFIG_PATA_LEGACY=n CONFIG_PATA_QDI=n CONFIG_PATA_WINBOND_VLB=n
to the base and PAE configs, and then the build continued. No idea if putting n was the right option, but I'm sure i'll find out when I run the kernel..
Hmm, that is odd. If you copy the config file from configs/ directory, make nonint_oldconfig should run without any needing for edit.
You would think that, but the configs in the config directory do not contain those options, and it wouldn't build without them.
either way the build has just failed with:
"make: *** [drivers] Error 2 + exit 1"
any ideas?
What is your uname -mr ? Also please show us a few lines above that error line in your error log.
Akemi
Akemi Yagi wrote
What is your uname -mr ? Also please show us a few lines above that error line in your error log.
I just reran the build changing those 4 config options to y, and it completed successfully. I'm running that kernel now, booted fine with no errors
Now I wait to see if the nfs problem is solved..
Thanks for you help
Dunc