Tom Diehl wrote: >On Sun, 20 Oct 2002, Randall J. Parr wrote: > > > >>I've installed RH8 (everything) and I have up2date'd to the most current >>patches. >> >>I am having a problem with rpm hanging when I run it from the command-line. >> >>I run an rpm command, for example, any of the following: >> >>---------------------------------------- >> >>rpm -qa | grep libdvd >> >>rpm -Uvh -vv --test libdvdcss*.rpm >LOG 2>&1 >> >>rpm -Uvh -vv --test apt* >LOG 2>&1 >> >>rpm -Uvh -vv LinNeighborhood* >LOG 2>&1 >> >>---------------------------------------- >> >>and 2/3 of the time the rpm process hangs, won't respond to a CTR-C or DEL >>and I have to issue a "kill -9 <pid>" to stop the hung process. >> >>I would figure my rpm db was hosed except that the up2date ran fine and >>1/3 of the time the rpm command runs fine and completes/installs just >>likes its supposed to. >> >>Is this a know problem or misconfiguration? >> >> > >Kinda!! See the rpm list archives for the gory details. Several things >can cause rpm to hang. The most common one is the presence of >/var/lib/rpm/__db* files. Remove them and run rpm -vv --rebuilddb and that >_should_ fix the problem. The other problem I am suddenly running into is >that rpm transactions such as installing or upgrading rpms on NFS mounted >file systems are failing. Copying the rpms onto the local machine allows >things to install/upgrade normally. I never had this problem before but it >is sure present now. The NFS problem is on a 7.2 system though so it >might not be relevant to this discussion. > >FYI anytime you abort an rpm installation or upgrade the __db* files will be >left around and should be removed. I have also seen several messages from >Jeff stating not to be too quick to abort an rpm transaction that appears >hung. He stated that if the rpm transaction is still using CPU cycles >it is most likely OK and to be patient. I cannot remember exactly what is >happening but I have seen this apparent hang symptom but the rpm transaction >was still using CPU cycles and eventually completed. If I had not looked >I would have aborted the transaction thinking it was hung. > >If the above does not fix your problem suggest you ask on the rpm list. Jeff >Johnson is very good at helping with these types of problems. > > >>This, and a few other things, has me very afraid of depending on RH8 at >>this point. >> I tried Tom's suggestions and was able to "unlock" rpm, that is, I could then run something like "rpm -qa | grep rhl" which worked fine. BUT rpm continues to hang. I thought maybe the problem was the rpms from freshrpms so I tried installing the Acrobat rpms for RH8 from GuruLabs acroread-5.06-4.i386.rpm acroread-plugin-5.06-4.i386.rpm rpm -K acroread* rpm -qa | grep acroread rpm -Uvh -vv --test acroread*.rpm >LOG 2>&1 works fine and LOG shows no dependency problem or other errors BUT rpm -Uvh -vv acroread*.rpm >LOG 2>&1 hangs and I have to kill -9 to cancel the rpm and then rm -f __db* to get it working again. This is REALLY becoming an show stopper for our use of RH8 From other posts I've seen on the rpm-list I may be experiencing a known rpm-hang problem. As per the other posts I have let it run a LONG time with no improvement. Nor have I been able to see in my LOG any consistency to the point it fails/hangs. Even worse if I try an install (like the above) several times it sometimes completes just fine. Is there a fix? Is this a problem with the rpms from freshrpms and the like? Any help would be greatly appreciated. R.Parr Temporal Arts -- Psyche-list mailing list Psyche-list@redhat.com https://listman.redhat.com/mailman/listinfo/psyche-list
