Dear ALUG,
Three things:
1) Is it possible to extract a single file from an RPM file (as if it were a tar archive, for instance)?
2) To ellaborate on 1): Last time I booted up my Mandake 9.1 box KDE (3.1) loaded without the kicker program (the panel at the bottom). Eventually I found that this was because /usr/lib/kicker.so was missing. I tried creating a symbolic link to /usr/lib/kicker.la but (as I half suspected) it didn't work (.so's and .la's clearly aren't that same thing!). So I found the kdebase.*.rpm on my discs and found that this rpm includes the kicker.so file but I can't find a way to get at that file without installing the entire package (which both I and Linux are reluctant to do!). Hence question 1).
3) I've had similar problems before and I believe I know the cause - but not the solution! Every 23rd boot my Mandrake box checks the root filesystem for errors and always finds and repairs them. On odd occasions after it's done this I get these sorts of problems - in fact its always been missing libraries.
So, if anyone has any answers to any of these questions I'd be very grateful to hear them!
Thanks very much,
Richard.
Linux forever, Windows never!
Richard Lewis Richard.Lewis@uea.ac.uk sed:
Dear ALUG,
Three things:
- Is it possible to extract a single file from an RPM file (as if it were a
tar archive, for instance)?
rpm2tgz.
Richard.
lewis
I'd also suggest that if your system is finding errors on / during the mount count fsck, then you have a far bigger problem than missing libs.
This check is done at boot after a preset number of mounts and providing that a fsck as a result of an unclean shutdown hasn't occurred during that count.
Unless there is something very wrong with your disks or filesystem it should not be finding any errors.
Backup your system and then check your disks (preferably with a surface scan) Always backup first if your suspect that there is a hardware error as the intensive operations performed during the check itself can finish off a dying disk.
On Wednesday 30 July 2003 19:27, Lewis Collard wrote:
Richard Lewis Richard.Lewis@uea.ac.uk sed:
Dear ALUG,
Three things:
- Is it possible to extract a single file from an RPM file (as if it
were a tar archive, for instance)?
rpm2tgz.
Richard.
lewis
On Wed, 30 Jul 2003 11:56:06 +0100 Richard Lewis Richard.Lewis@uea.ac.uk wrote:
Dear ALUG,
Three things:
- Is it possible to extract a single file from an RPM file (as if it were a
tar archive, for instance)?
I dont think so
- To ellaborate on 1): Last time I booted up my Mandake 9.1 box KDE (3.1)
loaded without the kicker program (the panel at the bottom). Eventually I found that this was because /usr/lib/kicker.so was missing. I tried creating a symbolic link to /usr/lib/kicker.la but (as I half suspected) it didn't work (.so's and .la's clearly aren't that same thing!). So I found the kdebase.*.rpm on my discs and found that this rpm includes the kicker.so file but I can't find a way to get at that file without installing the entire package (which both I and Linux are reluctant to do!). Hence question 1).
Dont worry do it anyway,
OR
#rpm --prefix /tmp/my_virtual_computer/root/ --dbpath /tmp/my_virtual_computer/pm_db/ -i --nodeps RPMS/i386/edg-se-libstringextra-cvs2003080103-1.i386.rpm
#ls /tmp/my_virtual_computer/root/lib/ libstringextra.a libstringextra.so libstringextra.so.0.0.0 libstringextra.la libstringextra.so.0
Its just a little C++ libarry of tools for me as I migrated myself away from Delphi to C++, BSD Licence.
- I've had similar problems before and I believe I know the cause - but not
the solution! Every 23rd boot my Mandrake box checks the root file system for errors and always finds and repairs them. On odd occasions after it's done this I get these sorts of problems - in fact its always been missing libraries.
Yes move to a journaled file system, you are loosing os files on occasion I bet you run ext2, best demonstrated using User Mode Linux.
Open files with a hard reboot will cause the system to delete the files.
I thought my data was safe on Linux but now I know better, I use ext3 but am considering either Reiserfs, or most likely xfs.
Thankyou SGI
So, if anyone has any answers to any of these questions I'd be very grateful to hear them!
Thanks very much,
Richard.
Linux forever, Windows never!
Regards
Owen
owen.churchcowley owen.churchcowley@ntlworld.com wrote:
I thought my data was safe on Linux but now I know better, I use ext3 but am considering either Reiserfs, or most likely xfs.
I've found that Reiserfs is still a work in progress and isn't that hard to cause data loss on, at least in 2.4.x -- I think others on #alug have said similar things, but I forget who.
----- Original Message ----- From: "MJ Ray" markj@cloaked.freeserve.co.uk To: main@lists.alug.org.uk Sent: Thursday, July 31, 2003 11:44 PM Subject: Re: [Alug]Re: Questions about rpm --prefix , a hard reboot will cause the ext2 system to delete the files, Delphi to C++ libary advert
owen.churchcowley owen.churchcowley@ntlworld.com wrote:
I thought my data was safe on Linux but now I know better, I use ext3 but am considering either Reiserfs, or most likely xfs.
I've found that Reiserfs is still a work in progress and isn't that hard to cause data loss on, at least in 2.4.x -- I think others on #alug have said similar things, but I forget who.
I used Reiserfs for a short while and had it do some strange things so I went back to ext2 and now to ext3 with no problems.
Cheers, BJ
I tried reiserfs a while back and quickly found I couldn't chop and change the hard drives between machines.. Switching to ext3, I found that if the drive started to die, it was still possible to recover data with a rescue system that only supported ext2.
The other advantage of ext3 filesystems is you can upgrade an ext2 partition without having to worry about reformatting first - Thus saving the chore of backing up first..
Regards, Paul.
On Thursday 31 July 2003 11:44 pm, MJ Ray wrote:
I've found that Reiserfs is still a work in progress and isn't that hard to cause data loss on, at least in 2.4.x -- I think others on #alug have said similar things, but I forget who.
The thing about Richard's problem is that it's not a result of hard shutdowns as he indicated that this happens at the mount count check not the check after a unclean unmount.
Also in my experience files that are only generally being read from (like his libs) are reasonably safe (even on ext2)
Journaling filesystems are more about protection in regards to writes.
I still think Richard's problem could be more to do with a hardware issue, in that case ext3 or reiser isn't going to save him. Richard, check your physical disks this to me smacks of a growing clump of bad sectors.
On Thursday 31 July 2003 22:44, MJ Ray wrote:
owen.churchcowley owen.churchcowley@ntlworld.com wrote:
I thought my data was safe on Linux but now I know better, I use ext3 but am considering either Reiserfs, or most likely xfs.
I've found that Reiserfs is still a work in progress and isn't that hard to cause data loss on, at least in 2.4.x -- I think others on #alug have said similar things, but I forget who.
It certainly is possible to extract one file from an rpm - I use Midnight Commander (mc) to browse the internal tree and just do a trivial copy.
Regards, Paul.
On Friday 01 August 2003 3:38 am, owen.churchcowley wrote:
1) Is it possible to extract a single file from an RPM file (as if it were a tar archive, for instance)?
I dont think so
-
John Woodard -
Lewis Collard -
MJ Ray -
owen.churchcowley -
Paul -
Richard Lewis -
Wayne Stallwood