I´m running the december raid patches on a lightly but constantly loaded
nfs soft-raid5 fileserver. Until now there wasn´t any problem during
normal operation, but I just got lots of:
"EXT2-fs error (device 09:00): ext2_new_block: Free blocks count
corrupted for block group 259"
in the log.
the effect:
There is a directory with about 5000 subdirectories, and in some of them
all file writes after create fail with "no space left on device".
Creating files works, the results are lots of 0 byte files. Other
subdirs work without problems. I had a tool delete all unused subdirs
(about 4500) and the problem disappeared (risky, admitted). I suppose I
should fsck the partition as soon as possible...
the reasons I can think of:
- ide drive problem
- ide cable problem (however: no udma, simple i/o polling)
- controller problem (promise)
- raid0145 problem
- ext2fs problem
did anyone else encounter similar problems?
ingo: did you fix anything since the december patch that might have been
a cause for this?
Is there a possibility to start the array check manually?
The system infos, syslog, dumpe2fs:
/proc/mdstat:
Personalities : [raid5]
read_ahead 1024 sectors
md0 : active raid5 hdh1[3] hdg1[2] hdf1[1] hde1[0] 49532736 blocks level
5, 32k chunk, algorithm 2 [4/4] [UUUU]
unused devices: <none>
uptime: 25 days
kernel 2.0.36
raid-patch: raid0145-19981215-2_0_36
other patches:
include/asm/param.h:
#define HZ 256 (orig: 100)
include/linux/fs.h:
#define NR_OPEN 1024 // 512 (orig: 256)
#define NR_INODE 12288 // 6144 (orig: 3072)
#define NR_FILE 4096 // 2048 (orig: 1024)
#define IS_NOATIME(inode) (MS_NOATIME)
#define UPDATE_ATIME(inode) { }
include/linux/limits.h:
#define NR_OPEN 1024 // 512 (orig: 256)
#define OPEN_MAX 1024 // 512 (orig: 256)
include/linux/tasks.h:
#define NR_TASKS 2048 // (orig: 512)
/var/log/messages:
Mar 6 02:47:28 fileserver kernel: EXT2-fs error (device 09:00):
ext2_new_block: Free blocks count corrupted for block group 259
Mar 6 02:47:31 fileserver kernel: EXT2-fs error (device 09:00):
ext2_new_block: Free blocks count corrupted for block group 259
Mar 6 09:31:46 fileserver kernel: EXT2-fs error (device 09:00):
ext2_new_block: Free blocks count corrupted for block group 259
Mar 6 09:31:58 fileserver last message repeated 7 times
Mar 6 10:13:09 fileserver kernel: EXT2-fs error (device 09:00):
ext2_new_block: Free blocks count corrupted for block group 259
Mar 6 10:44:29 fileserver kernel: EXT2-fs error (device 09:00):
ext2_new_block: Free blocks count corrupted for block group 259
Mar 6 10:45:05 fileserver last message repeated 74 times
Mar 6 10:48:26 fileserver kernel: EXT2-fs error (device 09:00):
ext2_new_block: Free blocks count corrupted for block group 259
Mar 6 11:26:30 fileserver kernel: EXT2-fs error (device 09:00):
ext2_new_block: Free blocks count corrupted for block group 259
Mar 6 11:26:59 fileserver last message repeated 64 times
(plus more of them)
dumpe2fs:
Filesystem volume name: <none>
Last mounted on: <not available>
Filesystem UUID: cb818caa-9adf-11d2-9ebf-006008bd9a0a
Filesystem magic number: 0xEF53
Filesystem revision #: 0 (original)
Filesystem features: (none)
Filesystem state: not clean with errors
Errors behavior: Continue
Filesystem OS type: Linux
Inode count: 1548288
Block count: 12383184
Reserved block count: 0
Free blocks: 2919245
Free inodes: 1526335
First block: 0
Block size: 4096
Fragment size: 4096
Blocks per group: 32768
Fragments per group: 32768
Inodes per group: 4096
Inode blocks per group: 128
Last mount time: Wed Feb 10 14:16:18 1999
Last write time: Sun Mar 7 17:24:15 1999
Mount count: 4
Maximum mount count: 20
Last checked: Sun Jan 17 11:19:13 1999
Check interval: 15552000 (6 months)
Next check after: Fri Jul 16 12:19:13 1999
Reserved blocks uid: 0 (user root)
Reserved blocks gid: 0 (group root)
