On Wed, Sep 10, 2003 at 12:19:27PM +0200, Stephan Sachse wrote: [incredibly useful stuff zapped]
> Herbert Poetzl <[EMAIL PROTECTED]> wrote: > > hmm, sounds interesting, what crashes did you observe? > here is the photo http://www.teg-clan.de/bilder/vserver-qouta-1.jpg hey, nice picture, (nice bookshelf too, by the way ;) could you please make an ascii copy of this oops and run it through ksymoops? (if you need help with ksymoops, let me know) > The system keeps running but all vserver* scripts hangs after the crash. > A reboot will fail because the vserver* scripts hangs only a hard reset > helps. > > > > i do the "chcontext --ctx 100 cqhadd -v /dev/hda3" after i start the > > > vServer. Now i must wroth a script that adds the quota before the > > > vServer starts. But the Problem is the /dev/real/block/device. How > > > can i find out die real block device like /dev/hda3 for a vServer on > > > demand when the vServer starts. > > > > well, if you know, you use /dev/hda3 for _all_ vservers, > > it would always be /dev/hda3 and further, you would only > > require _one_ vroot device, so setting up the vroot on > > boot time, and just using /dev/vroot/0 for each server > > would be sufficient, wouldn't it? > > i have defined a VARIABLE REALBLOCKDEVICE in v01.conf and handle the > vrsetup in /etc/init.d/vservers. The cqhadd/cqhrem is doing in v01.sh. > It's not so good because the REALBLOCKDEVICE is "hardcoded" in conf > file, but it works hmm, maybe I missed something, but for what exactly do you need the REALBLOCKDEVICE? best, Herbert > cya later > /stephan > > -- > "Perl - the only language that looks the same > before and after RSA encryption." -Keith Bostic
