My server froze during the night for some reason. I was anticipating fireworks upon reboot and I was not disappointed. The final drive took over 23 minutes to replay all the Reiserfs transactions. A new record for me, but I haven't had a "dirty shutdown" in over a year. What is the longest you've seen?

 

BTW, parity check hasn't found any errors yet - fingers crossed.

 

Oct 4 07:03:43 Tower kernel: REISERFS (device md4): replayed 232 transactions in 926 seconds
Oct 4 07:03:46 Tower kernel: REISERFS (device md4): Using r5 hash to sort names
Oct 4 07:05:37 Tower kernel: REISERFS (device md3): replayed 246 transactions in 1041 seconds
Oct 4 07:05:39 Tower kernel: REISERFS (device md3): Using r5 hash to sort names
Oct 4 07:05:55 Tower kernel: REISERFS (device md8): replayed 256 transactions in 1059 seconds
Oct 4 07:05:57 Tower kernel: REISERFS (device md8): Using r5 hash to sort names
Oct 4 07:07:41 Tower kernel: REISERFS (device md5): replayed 286 transactions in 1165 seconds
Oct 4 07:07:42 Tower kernel: REISERFS (device md5): Using r5 hash to sort names
Oct 4 07:08:09 Tower kernel: REISERFS (device md13): replayed 293 transactions in 1193 seconds
Oct 4 07:08:10 Tower kernel: REISERFS (device md13): Using r5 hash to sort names
Oct 4 07:08:13 Tower kernel: REISERFS (device md2): replayed 292 transactions in 1197 seconds
Oct 4 07:08:13 Tower kernel: REISERFS (device md2): Using r5 hash to sort names
Oct 4 07:08:45 Tower kernel: REISERFS (device md17): replayed 290 transactions in 1229 seconds
Oct 4 07:08:46 Tower kernel: REISERFS (device md17): Using r5 hash to sort names
Oct 4 07:08:48 Tower kernel: REISERFS (device md10): replayed 321 transactions in 1232 seconds
Oct 4 07:08:51 Tower kernel: REISERFS (device md10): Using r5 hash to sort names
Oct 4 07:08:51 Tower kernel: REISERFS (device md9): replayed 315 transactions in 1234 seconds
Oct 4 07:08:53 Tower kernel: REISERFS (device md9): Using r5 hash to sort names
Oct 4 07:09:26 Tower kernel: REISERFS (device md11): replayed 330 transactions in 1270 seconds
Oct 4 07:09:28 Tower kernel: REISERFS (device md11): Using r5 hash to sort names
Oct 4 07:09:37 Tower kernel: REISERFS (device md16): replayed 314 transactions in 1280 seconds
Oct 4 07:09:39 Tower kernel: REISERFS (device md16): Using r5 hash to sort names
Oct 4 07:09:53 Tower kernel: REISERFS (device md6): replayed 334 transactions in 1297 seconds
Oct 4 07:09:54 Tower kernel: REISERFS (device md6): Using r5 hash to sort names
Oct 4 07:10:04 Tower kernel: REISERFS (device md14): replayed 354 transactions in 1308 seconds
Oct 4 07:10:04 Tower kernel: REISERFS (device md14): Using r5 hash to sort names
Oct 4 07:10:09 Tower kernel: REISERFS (device md7): replayed 353 transactions in 1313 seconds
Oct 4 07:10:10 Tower kernel: REISERFS (device md7): Using r5 hash to sort names
Oct 4 07:10:14 Tower kernel: REISERFS (device md15): replayed 357 transactions in 1318 seconds
Oct 4 07:10:14 Tower kernel: REISERFS (device md15): Using r5 hash to sort names
Oct 4 07:10:32 Tower kernel: REISERFS (device md18): replayed 397 transactions in 1336 seconds
Oct 4 07:10:33 Tower kernel: REISERFS (device md18): Using r5 hash to sort names
Oct 4 07:10:46 Tower kernel: REISERFS (device md1): replayed 397 transactions in 1350 seconds
Oct 4 07:10:46 Tower kernel: REISERFS (device md1): Using r5 hash to sort names
Oct 4 07:11:39 Tower kernel: REISERFS (device md12): replayed 651 transactions in 1402 seconds

That seems like allot of unsync'ed data.

Was there stuff on the cache drive that was moved and yet not sync'ed?

I only use a cache drive with one of my data drives - which was one of the first to mount. I don't know why it had so many transactions to replay, but I'm glad it is such a robust file system.