Nate,
Thanks for investigating this. Sounds like ZFS is either conflicting
with the Linux partition or running off the end of its partition in the
VMware configuration you set up. The result is the CKSUM errors you
are observing. This could well lead to errors when we try to pagefault
in the iso image blocks at mount.
There is still a bug here I think in ZFS in the way it is handling these
pagefault errors. We should not be panicing.
Given your analysis, I don't think I need your crash dump.
Thanks for using (and finding a bug in) ZFS!
-Mark
Nathanael Burton wrote:
Mark,
I might know a little bit more about what's causing this particular panic. I'm
currently running OpenSolaris as a guest OS under VMware Server RC1 on a CentOS
4.3 host OS. I have 3 - 300GB (~280GB usable) SATA disks in the server that
are all formatted under CentOS like so:
[EMAIL PROTECTED] sdb]# fdisk -l /dev/sda
Disk /dev/sda: 300.0 GB, 300069052416 bytes
255 heads, 63 sectors/track, 36481 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System
/dev/sda1 * 1 3187 25599546 fd Linux raid autodetect
/dev/sda2 3188 36481 267434055 bf Solaris[/b]
So I use the first ~25GB/disk in a Linux software RAID 5, the rest of the disk ~240GB (usable) is given to OpenSolaris (via VMware) as a raw physical disk partition. OpenSolaris still thinks that those disks that it's been given are the full size (~280GB) -- PROBLEM 1.
Next I can create a simple ZFS pool using one of the SATA disks like so:
[b]bash-3.00# zpool create sata c2t0d0[/b]
Then I copy an iso file from my OpenBSD file server via ftp... As soon as data
starts writing into the ZFS file system I notice zpool CKSUM errors -- PROBLEM
2. The first time I saw this problem occur I never checked the output of zpool
status, and I believe I must have had a bunch of CKSUM errors then too.
Current info:
[b]bash-3.00# pwd
/data
bash-3.00# ls -al
total 1423398
drwxr-xr-x 2 root sys 3 Jun 15 20:55 .
drwxr-xr-x 43 root root 1024 Jun 15 20:57 ..
-rw-r--r-- 1 root root 728190976 Sep 23 2005
KNOPPIX_V4.0.2CD-2005-09-23-EN.iso
bash-3.00# zpool status
pool: sata
state: ONLINE
status: One or more devices has experienced an unrecoverable error. An
attempt was made to correct the error. Applications are unaffected.
action: Determine if the device needs to be replaced, and clear the errors
using 'zpool clear' or replace the device with 'zpool replace'.
see: http://www.sun.com/msg/ZFS-8000-9P
scrub: none requested
config:
NAME STATE READ WRITE CKSUM
sata ONLINE 0 0 20
c2t0d0 ONLINE 0 0 20
errors: No known data errors
bash-3.00# zfs list
NAME USED AVAIL REFER MOUNTPOINT
sata 695M 273G 24.5K /sata
sata/data 695M 273G 695M /data
sata/mp3s 24.5K 273G 24.5K /mp3s[/b]
Now, I attempt to mount the iso file via lofiadm and the panic occurs:
[b]bash-3.00# mount -F hsfs `lofiadm -a
/data/KNOPPIX_V4.0.2CD-2005-09-23-EN.iso` /tmp/test[/b]
I have also tested the above scenario but instead of giving OpenSolaris the SATA disk via raw physical disk access I create a VMware vmdk disk image file on the SATA disk and give that to OpenSolaris. In this case I can successfully create a ZFS file system, copy the same iso to it, and mount it via lofiadm.
So I have a new panic/crash dump -- it's absolutely huge, ~400MB after tar and bzip. If you still want it I can upload it to sunsolve as you requested. Or if there is a way to make it smaller let me know.
Thanks,
Nate
This message posted from opensolaris.org
_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
_______________________________________________
zfs-discuss mailing list
zfs-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
