LinuxQuestions.org
Help answer threads with 0 replies.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Newbie
User Name
Password
Linux - Newbie This Linux forum is for members that are new to Linux.
Just starting out and have a question? If it is not in the man pages or the how-to's this is the place!

Notices

Reply
 
Search this Thread
Old 02-08-2010, 06:50 PM   #1
kkutac001
LQ Newbie
 
Registered: Feb 2010
Location: San Antonio, Tx
Distribution: Fedora 9
Posts: 7

Rep: Reputation: 0
JFS partition full, though it should not be


Heya. I am not exactly new to linux, but I don't know a lot about filesystems, and my question is about one.

I have a 10gb partition I use for data. The /home is there, and I mount any other data partitions (like /music stuff) onto /data. These other mounted partitions add up to something like 60gb of diskspace, but since they're just mounted on /data, I believe they only take up 4096 bytes per mount point.

Some time ago, I found that the /data parition was full. There was only 330mb of data in /home, so I was perplexed. I found a cache dir under .opera that reported itself as having 132TB (yes, that is terrabytes) of files. I thought deleting the offending directory was the answer, so I deleted that cache dir and every file or subdirectory in it, but the /data partition is still like 99% full. I am a wee bit confused.

This very full /data partition is my only jfs partition. The other mounted filesystems are either ext3 or ntfs. Is it possible that the journal of this filesystem is corrupted? Or is hidden somewhere on the /data parition, taking up a bunch of space? (I obviously don't know enough about filesystem to know whether or not this is a likely scenario.) Is it possible to zero out (or delete and re-create) the journal, if so? The only other thing I can think of is to move all the /home data off, delete the partition, then re-create it and move /home back. I will do that if need be, but I'd rather learn something from the experience, weird as it is.

..kkutac
 
Old 02-08-2010, 07:45 PM   #2
mryuck
Member
 
Registered: Feb 2010
Posts: 107

Rep: Reputation: 23
Do df -h. What does it say for partition size.
 
Old 02-08-2010, 07:56 PM   #3
SharpyWarpy
Member
 
Registered: Feb 2003
Location: Florida
Distribution: Fedora 18
Posts: 862

Rep: Reputation: 90
Post the output of
fdisk -l
and the output of
df -hT
and we'll start there.
 
Old 02-08-2010, 07:59 PM   #4
AlucardZero
Senior Member
 
Registered: May 2006
Location: USA
Distribution: Debian
Posts: 4,647

Rep: Reputation: 524Reputation: 524Reputation: 524Reputation: 524Reputation: 524Reputation: 524
did you unmount and fsck it?
 
Old 02-08-2010, 09:03 PM   #5
kkutac001
LQ Newbie
 
Registered: Feb 2010
Location: San Antonio, Tx
Distribution: Fedora 9
Posts: 7

Original Poster
Rep: Reputation: 0
AlcuardZero -- yes, I did fsck it (using the appropriate jfs util, and of course after unmounting). It found a small error, a picayune number of bytes recovered.
Here is the output of the two other commands requested (/data is sda7):
~/> fdisk -l

Disk /dev/sda: 250.0 GB, 250059350016 bytes
255 heads, 63 sectors/track, 30401 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x0009b4e3

Device Boot Start End Blocks Id System
/dev/sda1 * 1 292 2345458+ 83 Linux
/dev/sda2 293 4865 36732622+ 5 Extended
/dev/sda5 293 1567 10241374+ 83 Linux
/dev/sda6 1568 1758 1534176 82 Linux swap / Solaris
/dev/sda7 1950 3224 10241406 83 Linux
/dev/sda8 1759 1949 1534176 82 Linux swap / Solaris

Partition table entries are not in disk order

Disk /dev/sdb: 500.1 GB, 500107862016 bytes
255 heads, 63 sectors/track, 60801 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x00036031

Device Boot Start End Blocks Id System
/dev/sdb1 1 1305 10482381 7 HPFS/NTFS
/dev/sdb2 1306 60801 477901620 5 Extended
/dev/sdb5 1306 6527 41945683+ 83 Linux
/dev/sdb6 6528 16848 82903401 83 Linux
/dev/sdb7 16849 19582 21960823+ 83 Linux
/dev/sdb8 19583 24804 41945683+ 83 Linux
/dev/sdb9 24805 27415 20972826 83 Linux
/dev/sdb10 27416 55481 225440113+ 83 Linux

======================================

~/> df -hT

Filesystem Type Size Used Avail Use% Mounted on
/dev/sda1 ext3 2.3G 1.5G 694M 68% /
/dev/sda5 ext3 9.7G 4.6G 4.6G 51% /usr
tmpfs tmpfs 756M 48K 756M 1% /dev/shm
gvfs-fuse-daemon
fuse.gvfs-fuse-daemon 2.3G 1.5G 694M 68% /root/.gvfs
/dev/sda7 jfs 9.8G 9.6G 226M 98% /data
/dev/sdb1 fuseblk 10G 8.6G 1.5G 86% /data/xpfad
/dev/sdb5 ext3 40G 5.1G 33G 14% /data/pix
/dev/sdb6 ext3 78G 7.3G 67G 10% /data/mus
/dev/sdb7 ext3 21G 6.1G 14G 31% /data/cds
/dev/sdb8 ext3 40G 494M 37G 2% /data/sys
/dev/sdb9 ext3 20G 173M 19G 1% /data/src
/dev/sdb10 ext3 212G 44G 158G 22% /data/bu
//kkhome/lhold
cifs 49G 4.5G 45G 10% /mnt/tmp
 
Old 02-09-2010, 12:17 AM   #6
chrism01
Guru
 
Registered: Aug 2004
Location: Sydney
Distribution: Centos 6.5, Centos 5.10
Posts: 16,287

Rep: Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034
Might be worth trying

df -i

might be (nearly) out of inodes...
Also

du -h

on the offending partition
 
Old 02-09-2010, 06:41 PM   #7
kkutac001
LQ Newbie
 
Registered: Feb 2010
Location: San Antonio, Tx
Distribution: Fedora 9
Posts: 7

Original Poster
Rep: Reputation: 0
Chrism01 - Yeah, I'd wondered about the inodes, too, but they look good (/data only has 7% in use):
Filesystem Inodes IUsed IFree IUse% Mounted on
/dev/sda1 146592 19557 127035 14% /
/dev/sda5 640848 243033 397815 38% /usr
tmpfs 193302 2 193300 1% /dev/shm
gvfs-fuse-daemon 146592 19557 127035 14% /root/.gvfs
/dev/sda7 500768 33002 467766 7% /data
/dev/sdb1 1573800 18169 1555631 2% /data/xpfad
/dev/sdb5 2624496 7413 2617083 1% /data/pix
/dev/sdb6 5175408 5732 5169676 1% /data/mus
/dev/sdb7 1373568 29 1373539 1% /data/cds
/dev/sdb8 2624496 4241 2620255 1% /data/sys
/dev/sdb9 1313280 13 1313267 1% /data/src
/dev/sdb10 14090240 114368 13975872 1% /data/bu
=============
the "du -h" command lists bajillions of files/dirs, including the files/dirs from the filesystems mounted ON /data. So everything stored on /dev/sdb1, /dev/sdb5, etc which are mounted on /data are shown. (The output of the command is obviously faaaaar too large to post here.)

So, I'm still stumped. Moving data off /data, deleting the partition, then re-creating still looks like my only option. But I'll still not know what caused this, nor how I can prevent its happening again. Any other ideas to debug my situation...?
 
Old 02-09-2010, 10:18 PM   #8
mryuck
Member
 
Registered: Feb 2010
Posts: 107

Rep: Reputation: 23
If /home is the only directory on /data thats not a mount point run du -h on /home. Check for anymore giant TB files.
I was expecting df to return the wrong partition size but thats not the case here.
 
Old 02-10-2010, 12:26 AM   #9
chrism01
Guru
 
Registered: Aug 2004
Location: Sydney
Distribution: Centos 6.5, Centos 5.10
Posts: 16,287

Rep: Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034
Just btw, is there any reason you've got 2 swap partitions? Maybe you could add one of them to the regular disk space (later).
 
Old 02-11-2010, 01:40 AM   #10
kkutac001
LQ Newbie
 
Registered: Feb 2010
Location: San Antonio, Tx
Distribution: Fedora 9
Posts: 7

Original Poster
Rep: Reputation: 0
mryuck -- while in /home, I issued this command:
~> find . -exec du -h {} \; | sort -nr | head -n20
I found twenty megabyte-sized files (some associated with games, some with ies4linux, some with .opera/opcache, some with openoffice). A 1mb file is not small, but certainly not huge enough to cause my full disk. All together they only add up to 300-some-odd mb.
chrism01 -- I purposely made the swap files 1.5mb each. I read somewhere that linux would only use 1.5mb of a given swap partition, and I wanted 3mb of swap. I can't remember where I read that, so I can't tell you the source of where I got that idea. But I nevertheless implemented it.
I have unpartitioned space on the harddrive, and could enlarge the /data partition, but that would not explain how I got to where I am. I was hoping that we might, if we figured it out, learn from whatever caused it. Why not try, you know what I mean? But maybe whatever happened is so left-field and rare that it's not worth spending the time to figure out.
*sigh* I suppose I should just go ahead and delete and re-create the /data partition. I will leave it as-is for a few more days, in case someone gets a brilliant idea for degugging, but then move forward.
Thanks for looking into it.
 
Old 02-11-2010, 05:53 PM   #11
chrism01
Guru
 
Registered: Aug 2004
Location: Sydney
Distribution: Centos 6.5, Centos 5.10
Posts: 16,287

Rep: Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034
Ahh, normally you'd want 1 swap partition (not swap file) approx 1 to 2x amt of RAM.
http://kbase.redhat.com/faq/docs/DOC-15252 as an example.
Linux can use swap files but that's not usual practice.
 
Old 02-11-2010, 09:10 PM   #12
kkutac001
LQ Newbie
 
Registered: Feb 2010
Location: San Antonio, Tx
Distribution: Fedora 9
Posts: 7

Original Poster
Rep: Reputation: 0
chrism01 -- Sorry, I wasn't precise with my wording. In my previous post, I used the words "swap files" and then later "partitions": I should have said "partitions" both times. So, to recap: I do not use swap files; I set up two partitions of 1.5gb each. I apologize for the confusion.
Nice link you provided, thanks for that. Truth to tell, Linux seems to be damn good at using RAM. I don't see the swap mem being used much. However, I have never looked at swap utilization during my rsnapshot operations; maybe it gets a little workout during those operations.
Best regards,
kkutac001
 
Old 02-11-2010, 10:01 PM   #13
chrism01
Guru
 
Registered: Aug 2004
Location: Sydney
Distribution: Centos 6.5, Centos 5.10
Posts: 16,287

Rep: Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034Reputation: 2034
Ok; the other thing that caught my eye was you specified your swaps as 1.5mb(!); now that's d*mn small, even for D*mn Small Linux
 
Old 02-11-2010, 10:23 PM   #14
kkutac001
LQ Newbie
 
Registered: Feb 2010
Location: San Antonio, Tx
Distribution: Fedora 9
Posts: 7

Original Poster
Rep: Reputation: 0
Arrgghh! Right again: I should have written GB, of course! ..kkutac
 
Old 02-12-2010, 08:05 PM   #15
mryuck
Member
 
Registered: Feb 2010
Posts: 107

Rep: Reputation: 23
have you tried unmounting the other partitions and checking df again? I know it shouldn`t matter but worth a shot
 
  


Reply

Tags
filesystem, jfs


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
accidentally deleted jfs partition berrance Linux - Software 7 01-06-2010 06:49 AM
General Partitioning Limits: How full is too full to partition? orangesky Linux - Hardware 1 12-05-2009 06:27 PM
Trying to rescue data from a corrupted jfs partition jchance Linux - Hardware 1 09-15-2006 12:27 PM
How do I convert an empty ext3 partition to a jfs partition? lothario Linux - Software 3 02-28-2006 10:40 AM
root partition gets full when using cp -a on another partition timsch75 Linux - General 2 10-02-2005 06:41 AM


All times are GMT -5. The time now is 01:12 PM.

Main Menu
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
identi.ca: @linuxquestions
Facebook: linuxquestions Google+: linuxquestions
Open Source Consulting | Domain Registration