Bug#918187: fai - hook error handling completely broken

2019-01-11 Thread Holger Levsen
On Fri, Jan 11, 2019 at 08:26:36PM +0100, Bastian Blank wrote: > On Fri, Jan 11, 2019 at 01:51:53PM +0100, Thomas Lange wrote: > > since you did not justified why this bug is grave, I will downgrade > > it to normal. > Well, I consider it RC for the cloud stuff as it makes image build silently >

Bug#918187: fai - hook error handling completely broken

2019-01-11 Thread Bastian Blank
On Fri, Jan 11, 2019 at 01:51:53PM +0100, Thomas Lange wrote: > since you did not justified why this bug is grave, I will downgrade > it to normal. Well, I consider it RC for the cloud stuff as it makes image build silently break. Do you really want to play this every time? > Did you tried to

Bug#918187: fai - hook error handling completely broken

2019-01-11 Thread Thomas Lange
> On Fri, 04 Jan 2019 06:07:38 +0100, Bastian Blank said: > Severity: grave Hi Bastian, since you did not justified why this bug is grave, I will downgrade it to normal. Did you tried to set STOP_ON_ERROR to a lower value? You are right the exit code of a hook can only be less than

Bug#918187: fai - hook error handling completely broken

2019-01-03 Thread Bastian Blank
Source: fai Version: 5.7.2 Severity: grave There seems to be no way to make hooks abort the run. I explicitely made a hook always fail and it always continued: | partition.GRUB_CLOUD_AMD64 FAILED with exit code 1. | Skipping task_partition | Calling hook: mountdisks.CLOUD | No fstab file