| View previous topic :: View next topic |
| Author |
Message |
senduran
n00b
Joined: 05 Aug 2005
Posts: 71
|
 Posted: Wed Aug 17, 2005 7:28 pm Post subject: Regular 'Remounting filesystem read-only' [solved] |
 |
|
I have a newly-setup 2.6.12-r6-gentoo-sources gentoo box with amd64 dual core on Asus A8V deluxe (via K8T800Pro, via VT8237) motherboard. I am not using the extra promise pdc20378 raid controller which is disabled in bios.
I have a new Seagate 160GB sata drive hosting OS and a new Seagate 400GB ide drive for data. Both have ext3 partitions that have begun to go read-only during extended bouts of writing (compiling on the OS drive, copying across ~300GB of data to the data drive). I don't think they're overheating - every time I check the data drive (eg. after completing an fsck) it is less than 55 degrees C (max should be 60). After unmounting an extended session of fsck is required to correct the numerous problems found (mostly multiply claimed blocks iirc).
The only kernal problem I know I'm suffering from is lost ticks - is there any chance lost ticks could mess up writes and the partitions? It certainly seems to me that I gain time at a greatly accelerated rate whilst copying to the data drive.
An example dmesg after a partition goes read-only:
| Code: |
EXT3-fs error (device hdc1): ext3_new_block: Allocating block in system zone - block = 67076224
Aborting journal on device hdc1.
EXT3-fs error (device hdc1) in ext3_reserve_inode_write: Journal has aborted
ext3_abort called.
EXT3-fs error (device hdc1): ext3_journal_start_sb: Detected aborted journal
Remounting filesystem read-only
__journal_remove_journal_head: freeing b_committed_data
|
In my kernal configuration I don't have generic/default IDE chipset support, only VIA82CXXX chipset support (and the things needed to display that option). DMA is on. Everything built into kernel.
Other than the regular loss of a filesystem or two the box seems fine, but obviously I don't enjoy sitting on this particular time bomb. Any suggestions please as to causes and/or possible work arounds or tests?
Last edited by senduran on Sat Aug 20, 2005 9:57 pm; edited 1 time in total |
|
| Back to top |
|
 |
moocha
Watchman
Joined: 21 Oct 2003
Posts: 5722
|
| Posted: Thu Aug 18, 2005 3:44 pm Post subject: |
|
|
I'd try a different kernel or the same drives on a different system. Then I'd emerge smartmontools and have the drives run extended offline tests (takes a few hours). And then I'd run badblocks for a few hours on it.
_________________
Military Commissions Act of 2006: http://tinyurl.com/jrcto
"Those who would give up essential liberty to purchase a little temporary safety deserve neither liberty nor safety."
-- attributed to Benjamin Franklin |
|
| Back to top |
|
|
senduran
n00b
Joined: 05 Aug 2005
Posts: 71
|
| Posted: Sat Aug 20, 2005 9:57 pm Post subject: |
|
|
I tried a different kernel - vanilla sources 2.6.13-rc6. I still get a lost ticks error but after a day of copying my clock only gained about a minute (compared to gaining around an hour per hour) and the partition didn't go read-only. Touch wood the problem is solved.
What am I missing by not having a gentoo sources kernel? Would it be worth trying the latest testing gentoo sources, or a gentoo sources 2.6.13 when one comes out? Any ideas on what it is about 2.6.12-r6-gentoo that causes the problem? |
|
| Back to top |
|
|
moocha
Watchman
Joined: 21 Oct 2003
Posts: 5722
|
| Posted: Sat Aug 20, 2005 11:17 pm Post subject: |
|
|
No idea. I'm running 2.6.11-hardened on servers and 2.6.9-gentoo on my desktops. I never upgrade the kernel once I found a stable and speedy one, except when there's some new functionality I really, really, really can't live without.
_________________
Military Commissions Act of 2006: http://tinyurl.com/jrcto
"Those who would give up essential liberty to purchase a little temporary safety deserve neither liberty nor safety."
-- attributed to Benjamin Franklin |
|
| Back to top |
|
|
|
Kernel & Hardware
