I've been running SuSE 9.1 (currently 2.6.5-7.111.5-default) on my Athlon XP 2600+ for 6+ months now. I started with a 120 GB drive that dual boots win2k and SuSE. A couple months back I added a 200 GB drive and partitioned it into a 140 GB NTFS and 60 GB ReiserFS.

df -h
Filesystem Size Used Avail Use% Mounted on
/dev/hda8 7.8G 5.7G 2.2G 73% /
tmpfs 253M 24K 253M 1% /dev/shm
/dev/hda1 7.9G 7.1G 808M 90% /windows/C
/dev/hda5 10G 2.8G 7.3G 28% /windows/D
/dev/hda6 88G 83G 5.1G 95% /windows/E
/dev/hdb5 140G 67G 74G 48% /windows/G

/dev/hdb1 is missing from above, as it currently isn't mounted as i'll explain below (it's usually mounted as /usr/hdb).

/etc/fstab
/dev/hda8 / reiserfs acl,user_xattr 1 1
/dev/hda7 swap swap pri=42 0 0
devpts /dev/pts devpts mode=0620,gid=5 0 0
proc /proc proc defaults 0 0
usbfs /proc/bus/usb usbfs noauto 0 0
sysfs /sys sysfs noauto 0 0
/dev/cdrecorder /media/cdrecorder subfs fs=cdfss,ro,procuid,nosuid,nodev,exec,iocharset=utf8 0 0
/dev/cdrom /media/cdrom subfs fs=cdfss,ro,procuid,nosuid,nodev,exec,iocharset=utf8 0 0
/dev/hdb1 /usr/hdb reiserfs acl,user_xattr 0 0
/dev/hda1 /windows/C ntfs ro,umask=0222 0 0
/dev/hda5 /windows/D vfat user,rw,exec,umask=000 0 0
/dev/hda6 /windows/E ntfs ro,umask=0222 0 0
/dev/hdb5 /windows/G ntfs ro,umask=0222 0 0

My 120 GB drive is running fine, but power outages keep taking the linux partition of my 200 GB drive out. The first time this occurred, I lost about 20% of my 400,000 files on that partition. After running "fsck.reiserfs --check /dev/hdb1" and "fsck.reiserfs --rebuild-tree /dev/hdb1" I got those 80% back... but the filenames were all nicely screwed up. Anyways, I recovered from that loss with some handy scripts and backups.

Yesterday, I had another power outage, and again lost the same partition. I'm pretty sure that I'm in the same boat, and will likely have to play the renaming game again.

This is unacceptable. I'm puzzled as my other ReiserFS partition is untouched by the outages. My plan is to format both drives, give linux the 120 GB drive and windows the 200 GB drive.

Whatever misconfiguration I have with that partition should be corrected by re-installing and re-partitioning. Can anyone recommend an alternative to save me the time and hassle? (can I find a way to compare /dev/hda8 and /dev/hdb1 to see how their configurations differ... YaST is useless for this, and i'm not familiar with partition-related commands in the shell).

Thanks,
Cory.

sfdisk -l

Disk /dev/hda: 238216 cylinders, 16 heads, 63 sectors/track
Warning: extended partition does not start at a cylinder boundary.
DOS and Linux will interpret the contents differently.
Warning: The partition table looks like it was made
for C/H/S=*/255/63 (instead of 238216/16/63).
For this listing I'll assume that geometry.
Units = cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0

Device Boot Start End #cyls #blocks Id System
/dev/hda1 * 0+ 1019 1020- 8193118+ 7 HPFS/NTFS
/dev/hda2 1020 14944 13925 111852562+ f W95 Ext'd (LBA)
/dev/hda3 0 - 0 0 0 Empty
/dev/hda4 0 - 0 0 0 Empty
/dev/hda5 1020+ 2324 1305- 10482381 c W95 FAT32 (LBA)
/dev/hda6 2325+ 13797 11473- 92156841 7 HPFS/NTFS
/dev/hda7 13798+ 13926 129- 1036161 82 Linux swap
/dev/hda8 13927+ 14944 1018- 8177053+ 83 Linux

Disk /dev/hdb: 24321 cylinders, 255 heads, 63 sectors/track
Units = cylinders of 8225280 bytes, blocks of 1024 bytes, counting from 0

Device Boot Start End #cyls #blocks Id System
/dev/hdb1 0+ 6079 6080- 48837568+ 83 Linux
/dev/hdb2 6080 24318 18239 146504767+ f W95 Ext'd (LBA)
/dev/hdb3 0 - 0 0 0 Empty
/dev/hdb4 0 - 0 0 0 Empty
/dev/hdb5 6080+ 24318 18239- 146504736 7 HPFS/NTFS

Linux is in the extended partition of the 120 GB drive (hda), and is the primary partition of the 200 GB drive (hdb)...

You should include notail in the fstab options if you are using LILO because LILO does screw up the directory tree with reiserfs partitions. Also if I was you, I would understand what acl and user_xattr mean and what they do to the filesystem.

Did you try mounting it as read-only and with the lazy mount option.

Get an in-line UPS. An in-line UPS is better than stand-by UPS. What an in-line UPS device does is it takes AC voltage and converts it to DC. The DC voltage charges and at the same time it powers the inverter. The inverter then converts DC to AC to be used by computers or other electronic equipment. If power goes out, the in-line UPS does not need to switch from AC to battery backup like stand-by UPS because the battery is in the same path all the time. UPS manufactures do not specify if the UPS devices is in-line or stand-by, so read the manual on their web site. In-line UPS devices are more expensive than stand-by.

Boot-loader is GRUB...

fsck.reiserfs --check /dev/hdb1
reiserfsck 3.6.13
...
Replaying journal..
Reiserfs journal '/dev/hdb1' in blocks [18..8211]: 0 transactions replayed
Checking internal tree../ 1 (of 3)/ 1 (of 94)/ 1 (of 86)block 622932: The level of the node (15636) is not correct, (1) expected
the problem in the internal node occured (622932), whole subtree is skipped
/ 2 (of 94)block 622934: The level of the node (14648) is not correct, (2) expected
the problem in the internal node occured (622934), whole subtree is skipped
/ 2 (of 3)/ 52 (of 122)/124 (of 166)block 622733: The level of the node (15272) is not correct, (1) expected
the problem in the internal node occured (622733), whole subtree is skipped
/ 53 (of 122)/ 2 (of 170)block 623151: The level of the node (16200) is not correct, (1) expected
the problem in the internal node occured (623151), whole subtree is skipped
/ 54 (of 122)/ 4 (of 168)block 626720: The level of the node (47148) is not correct, (1) expected
the problem in the internal node occured (626720), whole subtree is skipped
/ 55 (of 122)/ 2 (of 168)block 631925: The level of the node (17824) is not correct, (1) expected
the problem in the internal node occured (631925), whole subtree is skipped
/ 56 (of 122)/ 11 (of 168)block 634436: The level of the node (42972) is not correct, (1) expected
the problem in the internal node occured (634436), whole subtree is skipped
/ 57 (of 122)/ 2 (of 167)block 638864: The level of the node (48489) is not correct, (1) expected
the problem in the internal node occured (638864), whole subtree is skipped
/ 58 (of 122)/ 4 (of 169)block 644177: The level of the node (14785) is not correct, (1) expected
the problem in the internal node occured (644177), whole subtree is skipped
/ 59 (of 122)/ 1 (of 168)block 652110: The level of the node (18121) is not correct, (1) expected
the problem in the internal node occured (652110), whole subtree is skipped
/ 60 (of 122)block 652210: The level of the node (1519) is not correct, (2) expected
the problem in the internal node occured (652210), whole subtree is skipped
finished
Comparing bitmaps..vpf-10640: The on-disk and the correct bitmaps differs.
Bad nodes were found, Semantic pass skipped
11 found corruptions can be fixed only when running with --rebuild-tree
----

I can fix this... not a problem. Run rebuild-tree and I get most of my files back sans directory tree.

But how can I prevent this from happening again? Would Ext3 or something else be less vulnerable, or would just tuning the mount options prevent further damage?

Heeding your recommendation:
user_xattr = Enable user-specified extended attributes
acl = Enable Posix Access Control Lists

I don't think I need these, but i'm not certain... I suppose I could change fstab options to either:
defaults
or acl,user_xattr,notail

I've searched around a bit... but haven't found a comprehensive explanation of what user_xattr and acl actually DO to the file system...