Possible bug: ext3 misreporting filesystem usage
I've been getting a possible bug after running my system a few weeks. The ext3 partition's usage is being misreported. Right now, df -h says ive got no space left, but according to du /, I'm only using 17 gigs of my 40 gig drive. Restarting fixes the problem, so I'm thinking it might be some mis-handled variable in memory, not something on the disc itself? And, yes, I do know that du is right, not df, because I keep good track of my disc usage. This is pretty serious, it killed a 40+ hour process that i'll have to start over again from the beginning!
In article <[email protected]>,
Quoted 8 lines · Calvin Spealman
I've been getting a possible bug after running my system a few weeks. The ext3 partition's usage is being misreported. Right now, df -h says ive got no space left, but according to du /, I'm only using 17 gigs of my 40 gig drive. Restarting fixes the problem, so I'm thinking it might be some mis-handled variable in memory, not something on the disc itself? And, yes, I do know that du is right, not df, because I keep good track of my disc usage. This is pretty serious, it killed a 40+ hour process that i'll have to start over again from the beginning!
There's a process holding on to a 23 GB logfile that has been deleted. Try "ls -l /proc/*/fd/* 2>&1 | grep deleted" . Kill the process and you'll have your space back.
Mike.
Quoted 11 lines · Miquel van Smoorenburg
In article <[email protected]>,
Quoted 8 lines · Calvin Spealman
I've been getting a possible bug after running my system a few weeks. The ext3 partition's usage is being misreported. Right now, df -h says ive got no space left, but according to du /, I'm only using 17 gigs of my 40 gig drive. Restarting fixes the problem, so I'm thinking it might be some mis-handled variable in memory, not something on the disc itself? And, yes, I do know that du is right, not df, because I keep good track of my disc usage. This is pretty serious, it killed a 40+ hour process that i'll have to start over again from the beginning!
Quoted 5 lines
There's a process holding on to a 23 GB logfile that has been deleted. Try "ls -l
/proc/*/fd/*2>&1 | grep deleted" . Kill the process and you'll have your space back.Mike.
All that shows is a couple things from konq's http cache, nothing adding nearly to the 23 gigs.
If i delete some files, i have more space, but then the used space steadily increases until i have nothing left again. i am running a 2.6.6_rc1 kernel.