> On Wed, 2007-08-01 at 08:36 -0700, Gallie, Keith wrote: > > KG > Thank you. We've actually discovered mutiple rpm > databases on our > > systems... > > > > Seth > That's.... odd. > > > > You don't know the half of it... > > > > There is the venerable /var/lib/rpm > > > > kgallie 539> l /var/lib/rpm > > total 76868 > > -rw-r--r-- 1 root root 663552 Feb 14 17:03 __db.003 > > -rw-r--r-- 1 root root 1318912 Feb 14 17:03 __db.002 > > -rw-r--r-- 1 root root 16384 Feb 14 17:03 __db.001 > > -rw-r--r-- 1 rpm rpm 12288 Feb 20 06:38 Conflictname > > -rw-r--r-- 1 rpm rpm 12288 Aug 1 03:43 Triggername > > -rw-r--r-- 1 rpm rpm 86016 Aug 1 03:43 Sigmd5 > > -rw-r--r-- 1 rpm rpm 176128 Aug 1 03:43 Sha1header > > -rw-r--r-- 1 rpm rpm 208896 Aug 1 03:43 > Requireversion > > -rw-r--r-- 1 rpm rpm 393216 Aug 1 03:43 Requirename > > -rw-r--r-- 1 rpm rpm 12288 Aug 1 03:43 Pubkeys > > -rw-r--r-- 1 rpm rpm 114688 Aug 1 03:43 > Provideversion > > -rw-r--r-- 1 rpm rpm 663552 Aug 1 03:43 Providename > > -rw-r--r-- 1 rpm rpm 56676352 Aug 1 03:43 Packages > > -rw-r--r-- 1 rpm rpm 45056 Aug 1 03:43 Name > > -rw-r--r-- 1 rpm rpm 20480 Aug 1 03:43 Installtid > > -rw-r--r-- 1 rpm rpm 32768 Aug 1 03:43 Group > > -rw-r--r-- 1 rpm rpm 10514432 Aug 1 03:43 Filemd5s > > -rw-r--r-- 1 rpm rpm 1708032 Aug 1 03:43 Dirnames > > -rw-r--r-- 1 rpm rpm 10641408 Aug 1 03:43 Basenames > > this is fine - the __db files are portions of the locking > mechanism in rpm. They can be reasonably safely removed. > > > > > There is /usr/lib/rpmdb/x86_64-redhat-linux/redhat > > > > -rw-r--r-- 1 root root 24576 Jul 12 2006 Triggername > > -rw-r--r-- 1 root root 98304 Jul 12 2006 Sigmd5 > > -rw-r--r-- 1 root root 106496 Jul 12 2006 Sha1header > > -rw-r--r-- 1 root root 196608 Jul 12 2006 > Requireversion > > -rw-r--r-- 1 root root 368640 Jul 12 2006 Requirename > > -rw-r--r-- 1 root root 24576 Jul 12 2006 Pubkeys > > -rw-r--r-- 1 root root 106496 Jul 12 2006 > Provideversion > > -rw-r--r-- 1 root root 327680 Jul 12 2006 Providename > > -rw-r--r-- 1 root root 49152 Jul 12 2006 Name > > -rw-r--r-- 1 root root 32768 Jul 12 2006 Installtid > > -rw-r--r-- 1 root root 24576 Jul 12 2006 Group > > -rw-r--r-- 1 root root 10510336 Jul 12 2006 Filemd5s > > -rw-r--r-- 1 root root 1474560 Jul 12 2006 Dirnames > > -rw-r--r-- 1 root root 24576 Jul 12 2006 Conflictname > > -rw-r--r-- 1 root root 10534912 Jul 12 2006 Basenames > > -rw-r--r-- 1 root root 50298880 Jul 12 2006 Packages > > -rw------- 1 root root 663552 Aug 1 03:43 __db.003 > > -rw------- 1 root root 1318912 Aug 1 03:43 __db.002 > > -rw------- 1 root root 16384 Aug 1 03:43 __db.001 > > this is the rpmdb-redhat rpmdb package. This is the world's > most useless > item: > > rpm -e rpmdb-redhat > > it'll just go away > > > > What REALLY scares me is that the __db.00* files in the redhat > > directory have the same dates as all the other files in > /var/lib/tmp. Cool, huh? > > Somehow that just doesn't seem right... > > the __db files are just an artifact of odd db4 locking. But why do the dates for db locking files in one directory match the other directory? > > I decided to PUNT on that old crap and get some new crap. I have > > requested rpm 4.4.2.1 be loaded up and I'll have to hack on > yum 3.2.2 > > to remove the root user requirement. I know it's a little > late for the > > question but is that a good combo? > > umm - how do you expect to write to the fs and the system > rpmdb w/o root privs? To make yum happy uid = os.geteuid() lockfile = '/var/run/yum.pid' uid=0 lockfile='/home/kgallie/YUM/yum.pid' That IS the beauty of the whole thing. I don't want to write to the *system* rpmdbs. I want to write to project/package specific rpm dbs. Maybe a dozen lightweight repositories. Example, you are working on project x, project y and z exists but you don't care about those. You can examine project x installed stuff and available stuff and get what you want. I'm working on project x in Boston, someone in Sunnyvale release an update or something new that you may or may not want. You can go get it or not and not have to involve sys admin/IT. That is the model here. If you write it, you install it. > > Furthermore does python 2.4+ include 2.5? I've had some > python lawyers > > around here tell me "NO", the python 2.4+ means python > 2.4.* and does > > not include 2.5. > > that's correct. python 2.4 is not python 2.5 To be pedantic - does 2.4+ include 2.5. 2.5 had a nicer debugger. Speaking of which (I knoe this isn't a python mailing list) but what do y'all use for python debugging? Thanks _______________________________________________ Yum mailing list Yum@xxxxxxxxxxxxxxxxxxxx https://lists.dulug.duke.edu/mailman/listinfo/yum
