In the last month - definitely after going to 5.5 - I've tried to fsck a drive (340 days, or some such, unchecked). 960G RAID 5, I *think*, possibly serial port attachment to a JetStore RAID array. Every time I try, it gets to 70.0%, and stops. As in, I left it run last night, having started it late afternoon, and around 23:00, it was still exactly there, not even .1% more. On that, I also had a -dd flag, since running it the other day with a -d gave me nothing at all of debugging info; neither did the dd.
Note that my manager tried to run one on another server, also attached to a JetStore, and it also stopped at the exact same point. This has never happened before, and fsck has handled these in the past.
Anyone seen anything like this, and/or have ideas as to what I can do to fsck to at least *tell* me why it was gagging?
mark
On Tue, Jul 20, 2010 at 1:57 PM, m.roth@5-cent.us wrote:
Note that my manager tried to run one on another server, also attached to a JetStore, and it also stopped at the exact same point. This has never happened before, and fsck has handled these in the past.
Anyone seen anything like this, and/or have ideas as to what I can do to fsck to at least *tell* me why it was gagging?
Hmm.. my fsck uses 'V' rather than 'dd' for verbose..
Maybe try attaching with strace to see if it's still doing anything??
Kwan Lowe wrote:
On Tue, Jul 20, 2010 at 1:57 PM, m.roth@5-cent.us wrote:
Note that my manager tried to run one on another server, also attached to a JetStore, and it also stopped at the exact same point. This has never happened before, and fsck has handled these in the past.
Anyone seen anything like this, and/or have ideas as to what I can do to fsck to at least *tell* me why it was gagging?
Hmm.. my fsck uses 'V' rather than 'dd' for verbose..
Nahh... -d is debug; I tried that,and got nothing, then dd, hoping that would increase the verbosity of the debugging statements, of which there were none.
Maybe try attaching with strace to see if it's still doing anything??
I s'pose I could try it again tomorrow - it takes a while, and I don't want it running overnight, when it's doing backups. That's a thought.
mark
On 07/20/2010 10:57 AM, m.roth@5-cent.us wrote:
In the last month - definitely after going to 5.5 - I've tried to fsck a drive (340 days, or some such, unchecked). 960G RAID 5, I *think*, possibly serial port attachment to a JetStore RAID array. Every time I try, it gets to 70.0%, and stops. As in, I left it run last night, having started it late afternoon, and around 23:00, it was still exactly there, not even .1% more. On that, I also had a -dd flag, since running it the other day with a -d gave me nothing at all of debugging info; neither did the dd.
I've seen e2fsck hang on large arrays (terabyte range) before, particularly if you have lots of hard links. It's a bug in fsck.
http://www.redhat.com/archives/ext3-users/2007-March/msg00016.html
Benjamin Franz wrote:
On 07/20/2010 10:57 AM, m.roth@5-cent.us wrote:
In the last month - definitely after going to 5.5 - I've tried to fsck a drive (340 days, or some such, unchecked). 960G RAID 5, I *think*, possibly serial port attachment to a JetStore RAID array. Every time I try, it gets to 70.0%, and stops. As in, I left it run last night, having started it late afternoon, and around 23:00, it was still exactly there, not even .1% more. On that, I also had a -dd flag, since running it the other day with a -d gave me nothing at all of debugging info; neither did the dd.
I've seen e2fsck hang on large arrays (terabyte range) before, particularly if you have lots of hard links. It's a bug in fsck.
http://www.redhat.com/archives/ext3-users/2007-March/msg00016.html
ARGH! We're doing rsync backups on it, and use a *lot* of hard links.... Wonnaful, wonnaful.
mark
-
Benjamin Franz -
Kwan Lowe -
m.roth＠5-cent.us -
Whit Blauvelt