On 09/03/2015 21:25, Cantor, Scott wrote:

On 3/9/15, 10:15 AM, "Cantor, Scott" wrote:

...

I'm seeing an odd error on all my builds since last night on just a few of the platforms:

This seems to be intermittent, though happens more than not. I presume somehow it's depending on which system the build ends up submitted to run on, but the builds will sometimes work, but more often fail with the errors unpacking the cyrus-sasl package from the OS.

-- Scott

It depends if the build system runs xen or KVM. I notice this error happens only if the build system runs kvm. Some of my SLE11_SP2 and SUSE_SLE-11_SP1-standard packages are affected, any help would be greatly appreciable. Andreas talks about a broken initrd, please fix it. -- To unsubscribe, e-mail: opensuse-buildservice+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-buildservice+owner@opensuse.org

On 3/10/15, 3:43 AM, "Adrian Schröter" wrote:

Is this on our server? Do you have have some build logs then somewhere around?

It is on your server(s), and unfortunately unless there are historical logs kept, I was able to keep resubmitting builds until I randomly managed to get them all to complete. However once I complete my package testing I can probably use my home project to resubmit builds, have them fail, and then provide the log.

I checked that rpm on the server for i586 and x86_64 and it looks okay. But this is already the inner CentOS system. one would need to check if it is broken in some cache or if the centos rpm has a reason to yell here

If the architecture is that there are multiple slaves to run the build jobs, it seemed very much like some of them were affected and others not. -- Scott

On 10/03/2015 08:43, Adrian Schröter wrote:

On Montag, 9. März 2015, 14:15:40 wrote Cantor, Scott:

...

I'm seeing an odd error on all my builds since last night on just a few of the platforms:

The tail end of the CentOS 6 failure:

[ 27s] [37/229] preinstalled cracklib-dicts-2.8.16-4.el6 [ 27s] [38/229] preinstalled curl-7.19.7-26.el6_1.2 [ 27s] [39/229] preinstalled cvs-1.11.23-11.el6_0.1 [ 27s] [40/229] installing cyrus-sasl-2.1.23-13.el6 [ 27s] error: unpacking of archive failed on file /var/run/saslauthd: cpio: mkdir failed - No such file or directory [ 27s] exit ...

Is this on our server? Do you have have some build logs then somewhere around?

Here: https://build.opensuse.org/package/live_build_log/home:hamelg2/libpcap/SLE_1... -- To unsubscribe, e-mail: opensuse-buildservice+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-buildservice+owner@opensuse.org

On Montag, 16. März 2015, 20:52:52 wrote Gilles Hamel:

On 10/03/2015 08:43, Adrian Schröter wrote:

...

On Montag, 9. März 2015, 14:15:40 wrote Cantor, Scott:

...

I'm seeing an odd error on all my builds since last night on just a few of the platforms:

The tail end of the CentOS 6 failure:

[ 27s] [37/229] preinstalled cracklib-dicts-2.8.16-4.el6 [ 27s] [38/229] preinstalled curl-7.19.7-26.el6_1.2 [ 27s] [39/229] preinstalled cvs-1.11.23-11.el6_0.1 [ 27s] [40/229] installing cyrus-sasl-2.1.23-13.el6 [ 27s] error: unpacking of archive failed on file /var/run/saslauthd: cpio: mkdir failed - No such file or directory [ 27s] exit ...

Is this on our server? Do you have have some build logs then somewhere around?

Here: https://build.opensuse.org/package/live_build_log/home:hamelg2/libpcap/SLE_1...

thanks a lot for this example :) It is indeed a breakage caused by the initrd. It seems dracut/systemd initrd's are not usable anymore with non-systemd systems. I added a workaround hack for now to the build script until we find a proper way to fix this. => yet another reason that the build environment provides it's own kernel & initrd... -- Adrian Schroeter email: adrian@suse.de SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Jennifer Guild, Dilip Upmanyu, Graham Norton, HRB 21284 (AG Nürnberg) Maxfeldstraße 5 90409 Nürnberg Germany -- To unsubscribe, e-mail: opensuse-buildservice+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-buildservice+owner@opensuse.org

On Tuesday, March 17, 2015 08:26:28 AM Adrian Schröter wrote:

It is indeed a breakage caused by the initrd. It seems dracut/systemd initrd's are not usable anymore with non-systemd systems.

I added a workaround hack for now to the build script until we find a proper way to fix this.

=> yet another reason that the build environment provides it's own kernel & initrd...

I am not sure if this has anything to do with it, but it seems that since this afternoon, nothing builds against Fedora 21 anymore. The error that I get is : 22s] /usr/bin/qemu-kvm -no-reboot -nographic -vga none -net none -cpu host -kernel /boot/vmlinuz -initrd /boot/initrd -append root=/dev/sda panic=1 quiet no-kvmclock nmi_watchdog=0 rw rd.driver.pre=binfmt_misc elevator=noop console=ttyS0 init=/.build/build -m 3000 -drive file=/var/cache/obs/worker/ root_3/root,if=none,id=disk,serial=0,cache=unsafe -device ide-hd,drive=disk - drive file=/var/cache/obs/worker/root_3/ swap,if=none,id=swap,serial=1,cache=unsafe -device ide-hd,drive=swap -smp 4 [ 22s] qemu: error reading initrd /boot/initrd: No such file or directory No live log available: remote error: scons no logfile Which would be correct as that Fedora is not using an initrd, but calls it initramfs (which is actually the default naming from dracut). Regards Raymond -- To unsubscribe, e-mail: opensuse-buildservice+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-buildservice+owner@opensuse.org

On Donnerstag, 19. März 2015, 08:14:27 wrote Adrian Schröter:

On Mittwoch, 18. März 2015, 21:03:53 wrote Raymond Wooninck:

...

On Tuesday, March 17, 2015 08:26:28 AM Adrian Schröter wrote:

...

It is indeed a breakage caused by the initrd. It seems dracut/systemd initrd's are not usable anymore with non-systemd systems.

I added a workaround hack for now to the build script until we find a proper way to fix this.

=> yet another reason that the build environment provides it's own kernel & initrd...

I am not sure if this has anything to do with it, but it seems that since this afternoon, nothing builds against Fedora 21 anymore. The error that I get is :

22s] /usr/bin/qemu-kvm -no-reboot -nographic -vga none -net none -cpu host -kernel /boot/vmlinuz -initrd /boot/initrd -append root=/dev/sda panic=1 quiet no-kvmclock nmi_watchdog=0 rw rd.driver.pre=binfmt_misc elevator=noop console=ttyS0 init=/.build/build -m 3000 -drive file=/var/cache/obs/worker/ root_3/root,if=none,id=disk,serial=0,cache=unsafe -device ide-hd,drive=disk - drive file=/var/cache/obs/worker/root_3/ swap,if=none,id=swap,serial=1,cache=unsafe -device ide-hd,drive=swap -smp 4 [ 22s] qemu: error reading initrd /boot/initrd: No such file or directory No live log available: remote error: scons no logfile

Which would be correct as that Fedora is not using an initrd, but calls it initramfs (which is actually the default naming from dracut).

/boot/$kernel or $intrd means it tries to use the worker host one.

So, this is a broken build host here. But this error should be catched as "badhost" and the job should be re-assigned.

Is this the case?

yes, it is ...

If not, please give me a pointer to the build job (project etc)

The initrd creation on quite some workers is broken atm ... fixing ... -- Adrian Schroeter email: adrian@suse.de SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Jennifer Guild, Dilip Upmanyu, Graham Norton, HRB 21284 (AG Nürnberg) Maxfeldstraße 5 90409 Nürnberg Germany -- To unsubscribe, e-mail: opensuse-buildservice+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-buildservice+owner@opensuse.org

On 3/10/15, 4:20 AM, "Andreas Schwab" wrote:

"Cantor, Scott" writes:

...

[ 27s] error: unpacking of archive failed on file /var/run/saslauthd: cpio: mkdir failed - No such file or directory

The most probable cause is that /var/run is a dangling symlink.

Due to something I'm doing in my specfiles, or some kind of system issue? I've never seen anything like this before certainly in several years of using the service. -- Scott

On 3/13/15, 6:48 AM, "Adrian Schröter" wrote:

may be cause by broken packaging or by broken dependencies.

I red elsewhere that you suggessted that filesystem should be installed before sasl.

So I added this to CentOS-6 prjconf now:

Order: filesystem:cyrus-sasl

Some background, this is breaking dependency cycles of packages at defined places. The bug is still in the packages but the line works around it.

Thanks, I'll do some more testing. -- Scott

On 13/03/2015 11:48, Adrian Schröter wrote:

I red elsewhere that you suggessted that filesystem should be installed before sasl.

So I added this to CentOS-6 prjconf now:

Order: filesystem:cyrus-sasl

Some background, this is breaking dependency cycles of packages at defined places. The bug is still in the packages but the line works around it.

I have tried your workaround, and unfortunately it doesn't work here. I have also SLE11 builds failed with this error : error: unpacking of archive failed on file /var/run/dbus: cpio: mkdir failed - No such file or directory -- To unsubscribe, e-mail: opensuse-buildservice+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-buildservice+owner@opensuse.org