On Thu, 2007-02-08 at 09:29 -0500, Kevin H. Hobbs wrote: > On Wed, 2007-02-07 at 12:00 -0500, fedora-list-request@xxxxxxxxxx wrote: > > issues of rpm database corruption and yum-updatesd service running are > > separate except for yum-updatesd expects you to have a working rpm > > database and failing that, I can see where yum-updatesd is not gonna > > be > > happy (though it still shouldn't leak memory). > > I suspect that the leaking in yum-updatesd was the cause of the > corruption on my system. I know that the rpmdb corruption showed up > after the first time that I had to kill yum-updatesd. I realized that > there were updates that were not being applied, trued to run yum, I was > told that another yum was running, I waited a day, I killed yum-updatesd > which was already using RAM out of control, and then my RPM database was > broken. ---- well it would appear to be evident that the rpmdb corruption and yum-updatesd using lots of RAM were interrelated but not certain that this corruption was caused by yum-updatesd. It's possible I guess. Like all other daemon services, it is preferable to shut the service down... /sbin/service yum-updatesd stop than to issue a kill/pkill etc. to stop it because databases (be they sql or berkeley), cannot be certain to be left in a manageable state when a process is rudely stopped. ---- > > I've since repaired the database. I hope that means it's really > repaired. At least yum itself uses the database just fine. ---- As far as I can tell, there are two states for the rpmdb. Normal and broken. There is only one rpmdb and it is maintained, queried and updated in essentially the same manner whether you use rpm, yum commands, pup GUI interface, smart or the yum-updatesd daemon and whichever process you use to install updates should be of no consequence. ---- > > Indeed with FC-6 we have had rpm db corruption at the levels unseen > > since RHL 8.0 > > > > I see many people making this complaint on the list. > > > I would expect that an updated yum-updatesd on a system with a > > functioning rpm database should work as expected. > > > > I'd expect the same thing, do you have any suggestions for evidence that > I should collect to demonstrate that this isn't the case? I've obviously > convinced myself, but I'll give it another try, and collect a bit more > data. ---- I haven't searched through bugzilla for reports on yum-updatesd and I would guess that would be the best place to start. See if there's any reports, closed or still open and you can add your comments / results of your own testing there. Craig
