Xen devel - Jul 2010 - [PATCH] pyGrub: Implement error handling on kernel/initrd extraction

Hi,
this is the patch to implement error handling on kernel/initrd 
extraction to pyGrub. There was no error handling before my patch 
applied and this patch adds error handling code to kernel/initrd 
extraction for write code. When dom0 is running out of space the 
standard OSError is being printed to stderr and then pygrub exists when 
an error on writing those files occured.

This was tested on RHEL-5 dom0 with upstream Xen-4.1 unstable installed 
with RHEL-5 i386 PV guest for both cases when dom0 had enough space and 
when it didn''t. For the case where there were not enough space on dom0 
it returned the error "pyGrub: [Errno 28] No space left on device" and
then failed with "Boot loader didn''t return any data"
message. For dom0
with enough space the PV guest was started successfully.

Since we agreed that introduction of dom0-min-space is not the right way 
to go, it''s superseeded by this patch.

Michal

Signed-off-by: Michal Novotny <minovotn@redhat.com>

-- 
Michal Novotny<minovotn@redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat



_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Michal Novotny writes ("[Xen-devel] [PATCH] pyGrub: Implement error
handling on kernel/initrd extraction"):
> Since we agreed that introduction of dom0-min-space is not the right way 
> to go, it''s superseeded by this patch.
...
> -        os.write(tfd, data)
> -        os.close(tfd)
> +        try:
> +            os.write(tfd, data)
> +            os.close(tfd)
> +        except OSError, e:
> +            print >>sys.stderr, "pyGrub: %s" % str(e)
> +            sys.exit(1)
Thanks, but this is not even slightly correct.

Ian.

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On 07/26/2010 05:23 PM, Ian Jackson wrote:
> Michal Novotny writes ("[Xen-devel] [PATCH] pyGrub: Implement error
handling on kernel/initrd extraction"):
>    
>> Since we agreed that introduction of dom0-min-space is not the right
way
>> to go, it''s superseeded by this patch.
>>      
> ...
>    
>> -        os.write(tfd, data)
>> -        os.close(tfd)
>> +        try:
>> +            os.write(tfd, data)
>> +            os.close(tfd)
>> +        except OSError, e:
>> +            print>>sys.stderr, "pyGrub: %s" % str(e)
>> +            sys.exit(1)
>>      
> Thanks, but this is not even slightly correct.
>
> Ian.
>    
Why not? It''s been tested and working fine. The try/except block is 
working fine and sys.exit(1) is necessary to terminate pyGrub.

Michal

-- 
Michal Novotny<minovotn@redhat.com>, RHCE
Virtualization Team (xen userspace), Red Hat


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On 07/27/2010 10:49 AM, Michal Novotny wrote:
> On 07/26/2010 05:23 PM, Ian Jackson wrote:
>> Michal Novotny writes ("[Xen-devel] [PATCH] pyGrub: Implement
error
>> handling on kernel/initrd extraction"):
>>> Since we agreed that introduction of dom0-min-space is not the
right way
>>> to go, it''s superseeded by this patch.
>> ...
>>> - os.write(tfd, data)
>>> - os.close(tfd)
>>> + try:
>>> + os.write(tfd, data)
>>> + os.close(tfd)
>>> + except OSError, e:
>>> + print>>sys.stderr, "pyGrub: %s" % str(e)
>>> + sys.exit(1)
>> Thanks, but this is not even slightly correct.
>>
>> Ian.
>
> Why not? It''s been tested and working fine. The try/except block
is
> working fine and sys.exit(1) is necessary to terminate pyGrub.
If os.write gets the OSError it will write the error message as part of 
the backtrace, and exit anyway.

That said I could reproduce the failure Michal started from, which is 
this error:

Error creating domain: (1, ''Internal error'',
''xc_dom_do_gunzip: inflate
failed (rc=-5)\\n'')

But I think xend can be forgiven for not treating very well about 
out-of-disk-space situations... let''s just not care.

Paolo

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel