dnf doesn't check for rpmdb lock?

Christopher ctubbsii-fedora at apache.org
Thu Jul 16 18:20:42 UTC 2015


On Thu, Jul 16, 2015 at 1:42 PM, Przemek Klosowski
<przemek.klosowski at nist.gov> wrote:
> On 07/16/2015 01:19 PM, Eric Griffith wrote:
>
> I tried to do an install and an update on two different terminals on my
> machine yesterday. The second one didn't yell about an rpmdb lock but it did
> say that it was waiting on a process. So unless it broke from an update last
> night / this morning then idk
>
> On Jul 16, 2015 11:20 AM, "Richard Shaw" <hobbes1069 at gmail.com> wrote:
>>
>> I remember frequently (to my dismay) getting messages from yum saying that
>> another process had a lock on the RPM database.
>>
>> Due to a recent rash of issues with the akmods package I have been
>> investigating why the kmods are not getting installed.
>>
>> Out of curiosity I tried running two dnf update instances in two different
>> terminals and I did not get the message, in fact it acted like nothing was
>> wrong!
>>
>
>
> So I just did 'dnf update' from two different terminals on the same system
> and allowed the first one to proceed. While it was updating, I kept running
> 'dnf update' in the other terminal, terminating it by not giving it an OK to
> proceed.
>
> The second 'dnf update'  was running, reporting decreasing number of
> upgrades needed, as expected. I don't know what would happen if I allowed
> the second one to proceed---would they mess up the package database? would
> two upgrades collide writing to the same packaged files?
>

I believe I tried this (accidentally) the other day, and the second
one just quit with an error message about another process. In yum, it
would keep retrying, but dnf just seems to give up on the first
collision.


More information about the devel mailing list