0.50rc1 feedback

[Jeff Johnson]

On Dec 8, 2006, at 1:36 PM, Neal Becker wrote:

> On Friday 08 December 2006 8:56 am, Gustavo Niemeyer wrote:
>> Ok folks,
>>
>> How's 0.50rc1 working for you?
>>
>> Have you found any issues to be fixed before 0.50 final, or is it
>> ok to relable 0.50rc1 as 0.50?
>
> I have been having some serious problems here that I haven't quite  
> figured
> out.  I run ksmarttray at startup, and manually run smart as needed.
> Approximately once every 1-2 days I run smart only to have it  
> hang.  I have
> to kill all smart instances (ksmarttray may be running 1), and
> find /var/lib/rpm/__* is trashed.  I have to rm them, run rpm -- 
> rebuilddb,
> which takes a LONG time on this machine.
>
> I don't know what's causing this, or if it is related to 0.50rc1.   
> It is
> relatively recent that I started running ksmarttray at startup.   
> Could it be
> an issue of locking when ksmarttray runs smart while I am trying to  
> run
> manually?

Is this FC6? If so, the problems are likely due to changes with yum/ 
pirut.

FWIW, rpm-4.4.8-0.4 uses robust posix mutexes, detects/purges stale  
read locks,
and detects (an likely corrects all but the worst cases) of write  
locks leaving
crud in database files.

In other words: No hangs, no --rebuilddb.

73 de Jeff