Re: [Cooker] This urpmi problem really drives me nuts
On Fri, Oct 05, 2001 at 04:50:19PM +0400, Borsenkow Andrej wrote: [root@cooker msec]# ll /var/lib/urpmi total 16 -rw-r--r--1 root root13973 Sep 30 22:48 compssUsers.flat [root@cooker msec]# urpmi.addmedia cooker file://cooker/i586/Mandrake/RPMS with ../base/hdlist.cz reading hdlist file [/var/lib/urpmi/hdlist.cooker.cz] built hdlist synthesis file for medium cooker found 0 headers in cache removing 0 obsolete headers in cache write config file [/etc/urpmi/urpmi.cfg] [root@cooker msec]# ll /var/lib/urpmi total 13224 -rw-r--r--1 root root40039 Oct 5 16:44 compss -rw-r--r--1 root root13973 Sep 30 22:48 compssUsers.flat -rw-r--r--1 root root 195803 Oct 5 16:44 depslist.ordered -rw-r--r--1 1137 1100 12725363 Oct 5 16:35 hdlist.cooker.cz -rw---1 root root 132649 Oct 5 16:44 list.cooker -rw-r--r--1 root root 320948 Oct 5 16:44 provides -rw-r--r--1 root root98654 Oct 5 16:44 synthesis.hdlist.cooker.cz [root@cooker msec]# urpmi.addmedia contribs file://cooker/i586/Mandrake/RPMS2 wi th ../base/hdlist2.cz reading hdlist file [/var/lib/urpmi/hdlist.contribs.cz] built hdlist synthesis file for medium contribs reading hdlist file [/var/lib/urpmi/hdlist.cooker.cz] reading hdlist file [/var/lib/urpmi/hdlist.contribs.cz] keeping only files referenced in provides reading hdlist file [/var/lib/urpmi/hdlist.cooker.cz] reading hdlist file [/var/lib/urpmi/hdlist.contribs.cz] computing dependencies write depslist file [/var/lib/urpmi/depslist.ordered] write provides file [/var/lib/urpmi/provides] write compss file [/var/lib/urpmi/compss] found 0 headers in cache removing 0 obsolete headers in cache write config file [/etc/urpmi/urpmi.cfg] [root@cooker msec]# ll /var/lib/urpmi total 16648 -rw-r--r--1 root root0 Oct 5 16:47 compss -rw-r--r--1 root root13973 Sep 30 22:48 compssUsers.flat -rw-r--r--1 root root0 Oct 5 16:47 depslist.ordered This is certainly the problem. urpmi cannot work well if there'are no packages in depslist.ordered. -rw-r--r--1 1137 1100 3944662 Oct 5 16:35 hdlist.contribs.cz -rw-r--r--1 1137 1100 12725363 Oct 5 16:35 hdlist.cooker.cz -rw---1 root root72405 Oct 5 16:46 list.contribs -rw---1 root root 132649 Oct 5 16:44 list.cooker -rw-r--r--1 root root 1075 Oct 5 16:47 provides -rw-r--r--1 root root40938 Oct 5 16:46 synthesis.hdlist.contribs.cz -rw-r--r--1 root root98654 Oct 5 16:44 synthesis.hdlist.cooker.cz [root@cooker msec]# urpmi diald no package named diald Shit! [root@cooker msec]# rpm -q urpmi rpmtools urpmi-1.7-14mdk rpmtools-3.1-4mdk -- [EMAIL PROTECTED]
RE: [Cooker] This urpmi problem really drives me nuts
[root@cooker msec]# urpmi diald no package named diald Shit! Can you try urpmi.update -a, Mostly the same effect (I get zero size depslist) if problem persist, try with urpmi.update -f -a Still there if problem still persist. Check access right to directory As root?! [root@cooker i386-linux]# ll -d /var/lib/urpmi /var/cache/urpmi drwxr-xr-x5 root root 120 Sep 24 20:21 /var/cache/urpmi/ drwxr-xr-x2 root root 272 Oct 5 17:47 /var/lib/urpmi/ and available size on it (/var/lib/urpmi as well as /var/cache/urpmi). urpmi need /var/lib and /var/cache on the same partition. I have just one Filesystem 1k-blocks Used Available Use% Mounted on /dev/hda5 3582344 1166272 2416072 33% / none 14788 0 14788 0% /dev/shm /dev/hda7 4008380 2670524 1134236 71% /mnt/hd I should add that this system has 32MB RAM. If it makes any difference. -andrej
Re: [Cooker] This urpmi problem really drives me nuts
Franc,ois Pons wrote: Borsenkow Andrej [EMAIL PROTECTED] writes: I should add that this system has 32MB RAM. If it makes any difference. Yes, but you have added swap, so it shouldn't matter. Urpmi and other processes may take up to 80 MB with a relatively complex urpmi database to build dependancies, so if you have at least 64 MB of swap, it may do it. I have 256MB of swap and I checked when urpmi was running, there was still free memory and plenty of swap. Still, here @home I have never seen this problem and I update from the same source (albeit not so frequently). So it may be related to memory size somehow. -andrej