Re: 3rd Fedora disk curdle
Don't know if it relates, but I once had ext3 curdling caused by a system firmware defect that changed how the BIOS reported disk sizes to the kernel ... it would work fine provided I always booted cold, inclusive-or never used a particular video mode. Once in the state, the reported disk size was different, partition alignment changed, and fsck kicked the dunny down. -- James Cameronmailto:[EMAIL PROTECTED] http://quozl.netrek.org/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: 3rd Fedora disk curdle
M. Edward (Ed) Borasky wrote: > Yeah, VMware can't deal very well with host-side platform problems. :) > > How sure are you that VMware Server 1.04 works on Windows XP 64 SP2? I'm > running VMware Workstation *6* on my AMD64 box -- I haven't messed with > the free server in a while and I've never tried it on a 64-bit machine. > Weelll, the site states that it works for 64bit clients and hosts and the install did not balk when I lit it up the XP64 machine so I assume I am ok. If I do not get problems with the VM on the direct connect IDE. I may take the time to build a SCSI based VM planted on the IDE drive and see if that invokes the problem. Ever see a compiler go crazy on the old HP mini's running HPDOS? You know, the OS that had the compilers own a raw piece of disk for the object code, so they did physical IO to the platter for compiles? That must have been about a ga-zillion years ago :) ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: 3rd Fedora disk curdle
Gerard J. Cerchio wrote: > M. Edward (Ed) Borasky wrote: >> 1. Define "curdle their ext3 disk". >> > I use the word curdle to describe a disk with lost inodes, sectors that > are multiply allocated, and other such problems that fsck valiantly > tries to correct but winds up with a non-working system. In the last > case I got a clean boot and clean subsequent fsck, however many > applications just seg faulted when I tried to run them. >> 2. What version of VMware are you using? >> > VMWare server 1.04 (latest freebe) >> 3. What's your host version? >> > Windows XP 64 SP2. Booting off an Nvidia RAID5 4 disk 1 TB. >> 4. Are you using the default settings on your virtual disks or are you >> changing them in some way? >> > The first two I just took all the defaults. This new one I put on the > IDE controller. I don't have a whole bunch of faith in the snapshot > system either so I just don't use it. > > Now here is something interesting, I just did a NTFS chkdsk on the RAID > and it moved my vmdk's into a copy of the the VM's directory while > reporting 2 lost files. These are the first errors I have ever seen on > the NVidia RAID after running it for a year. This has me nervous enough > to move the VM's off the RAID and onto an IDE backup drive in the > system. Maybe I am looking at some VMare/XP64 interaction. Anyway I'll > report any new problems on this thread. Yeah, VMware can't deal very well with host-side platform problems. :) How sure are you that VMware Server 1.04 works on Windows XP 64 SP2? I'm running VMware Workstation *6* on my AMD64 box -- I haven't messed with the free server in a while and I've never tried it on a 64-bit machine. ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: 3rd Fedora disk curdle
This reminds me of a situation I ran into about a zillion years ago, using V6 Unix: The filesystem and the swapper disagreed about the boundary between the FS and swap areas, so parts of the FS were getting swapped onto. I suppose something like that might be possible with certain pathological partition map inconsistencies. ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: 3rd Fedora disk curdle
M. Edward (Ed) Borasky wrote: > 1. Define "curdle their ext3 disk". > I use the word curdle to describe a disk with lost inodes, sectors that are multiply allocated, and other such problems that fsck valiantly tries to correct but winds up with a non-working system. In the last case I got a clean boot and clean subsequent fsck, however many applications just seg faulted when I tried to run them. > 2. What version of VMware are you using? > VMWare server 1.04 (latest freebe) > 3. What's your host version? > Windows XP 64 SP2. Booting off an Nvidia RAID5 4 disk 1 TB. > 4. Are you using the default settings on your virtual disks or are you > changing them in some way? > The first two I just took all the defaults. This new one I put on the IDE controller. I don't have a whole bunch of faith in the snapshot system either so I just don't use it. Now here is something interesting, I just did a NTFS chkdsk on the RAID and it moved my vmdk's into a copy of the the VM's directory while reporting 2 lost files. These are the first errors I have ever seen on the NVidia RAID after running it for a year. This has me nervous enough to move the VM's off the RAID and onto an IDE backup drive in the system. Maybe I am looking at some VMare/XP64 interaction. Anyway I'll report any new problems on this thread. ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: 3rd Fedora disk curdle
Gerard J. Cerchio wrote: > In the short time I have been working with olpc I have had my Fedora > VMware machines curdle their ext3 disks 3 times. > > I have been running 2.4 and 2.6 Redhats and Debians for over a year with > no such problems. Once the first Fedora 7 machine broke its disk I have > been very careful to shutdown every time. I cannot point to any > particular activity that has corrupted the disk image. Things will start > to go "wrong" and a subsequent reboot with fschk will yield a disk > hopelessly in trouble. > > Does the jhbuild emulator do any exotic direct to disk IO that may be > causing this? > Does Fedora aggressively modify its ext3, vfs or SCSI drivers? > > I have built a third Fedora 8 VM this time using IDE disk IO in hopes > that this problem will go away. > > Has anyone else seen this kind of problem? 1. Define "curdle their ext3 disk". 2. What version of VMware are you using? 3. What's your host version? 4. Are you using the default settings on your virtual disks or are you changing them in some way? My answers to 2 - 4: 2. VMware Workstation 6 3. Gentoo Linux on an AMD64 (Athlon64 X2 dual-core with 4 GB of RAM) 4. No ... I turn off the snapshots and make my virtual disks independent and persistent, and I turn on write caching. I have built virtual XOs, virtual Fedora 7 and 8 systems (32-bit!!), and jhbuild inside a virtual Fedora 8 system, all without any data corruption. The XOs I build with IDE drives because they won't boot from a SCSI drive -- "hda" is hard-wired somewhere that I don't recall off the top of my head -- but all the others use SCSI drives with no problem. So the first thing I recommend is to check your hard disk settings in VMware. The defaults *should* work without incident, but the snapshot logic may be doing something to you. > > -Gerard > ___ > Devel mailing list > Devel@lists.laptop.org > http://lists.laptop.org/listinfo/devel > ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: 3rd Fedora disk curdle
Gerard J. Cerchio wrote: > Does the jhbuild emulator do any exotic direct to disk IO that may be > causing this? Not that I know of. > Does Fedora aggressively modify its ext3, vfs or SCSI drivers? The Fedora kernel is very close to the upstream. > Has anyone else seen this kind of problem? Neither in real systems, nor in QEMU. -- \___/ |___| Bernardo Innocenti - http://www.codewiz.org/ \___\ One Laptop Per Child - http://www.laptop.org/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
3rd Fedora disk curdle
In the short time I have been working with olpc I have had my Fedora VMware machines curdle their ext3 disks 3 times. I have been running 2.4 and 2.6 Redhats and Debians for over a year with no such problems. Once the first Fedora 7 machine broke its disk I have been very careful to shutdown every time. I cannot point to any particular activity that has corrupted the disk image. Things will start to go "wrong" and a subsequent reboot with fschk will yield a disk hopelessly in trouble. Does the jhbuild emulator do any exotic direct to disk IO that may be causing this? Does Fedora aggressively modify its ext3, vfs or SCSI drivers? I have built a third Fedora 8 VM this time using IDE disk IO in hopes that this problem will go away. Has anyone else seen this kind of problem? -Gerard ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
3rd Fedora disk curdle
In the short time I have been working with olpc I have had my Fedora VMware machines curdle their ext3 disks 3 times. I have been running 2.4 and 2.6 Redhats and Debians for over a year with no such problems. Once the first Fedora 7 machine broke its disk I have been very careful to shutdown every time. I cannot point to any particular activity that has corrupted the disk image. Things will start to go "wrong" and a subsequent reboot with fschk will yield a disk hopelessly in trouble. Does the jhbuild emulator do any exotic direct to disk IO that may be causing this? Does Fedora aggressively modify its ext3, vfs or SCSI drivers? I have built a third Fedora 8 VM this time using IDE disk IO in hopes that this problem will go away. Has anyone else seen this kind of problem? -Gerard ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel