On 15/12/11 00:01, Akemi Yagi wrote:
> On Wed, Dec 14, 2011 at 2:06 PM, Farkas Levente<[log in to unmask]>  wrote:
>> On 12/14/2011 08:51 PM, Pat Riehecky wrote:
>>> Updated packages for nfs-utils posted for testing at
>>>
>>> ftp://ftp.scientificlinux.org/linux/scientific/6rolling/testing/i386/
>>> ftp://ftp.scientificlinux.org/linux/scientific/6rolling/testing/x86_64/
>>>
>>> These packages should not segfault in the manner that the current
>>> packages do.  If it does continue to segfault, or operate in an
>>> unexpected manner, please let me know.
>>>
>>> To test these you will need to downgrade to a version earlier than
>>> nfs-utils-1.2.3-15.el6 (yum downgrade nfs-utils), enable the testing
>>> repository, and install nfs-utils from there.  Once installed you can
>>> disable the testing repo.
>>>
>>> Or you can just run the following commands
>>>
>>> yum -y downgrade nfs-utils
>>> yum -y --enablerepo=sl-testing update nfs-utils
>>
>> is it an upstream fix or sl modified one?
>
> This SL post/thread has more details:
>
> http://listserv.fnal.gov/scripts/wa.exe?A2=ind1112&L=scientific-linux-users&T=0&P=16967

Two questions:

1) What actually caused the corrupted package build, and how sure are we 
no other packages have suffered similar problems?

2) Although I realize SL keeps the evr of TUV's rpm the same, in this 
case (post-SL release) I think it might be a good idea to bump the 
release (adding a .1 to the very end) to enable easy fixing of machines 
with the broken packages installed by a simple yum update. Thoughts?

Cheers,
Jonathan