No room at the Inn

[jorj]

I have been in the hospital for a month recoverering from pneumonia - no fun at 81.
My computer was on the whole time collecting junk mail and a few valid messages. I tried to do an apt-get update and I get the following message:

Hit ftp://ftp.debian.org sarge/non-free Release
Get:4 http://www.backports.org sarge-backports/main Release [137B]
Get:5 http://ftp.us.debian.org etch/main Release [81B]
Get:6 http://ftp.us.debian.org etch/contrib Packages [72.6kB]
Err http://ftp.us.debian.org etch/contrib Packages
Error writing to output file - write (28 No space left on device) [IP: 35.9.37.225 80]
Err http://ftp.us.debian.org etch/contrib Release
Bad header line [IP: 35.9.37.225 80]
98% [1 Packages gzip 5427200] [Waiting for headers]

also I checked the partition usage:

[root@speedy:/home/jorj]# df -h /dev/hda1
Filesystem Size Used Avail Use% Mounted on
/dev/hda1 7.9G 7.9G 0 100% /
[root@speedy:/home/jorj]# df -h /dev/hda2
Filesystem Size Used Avail Use% Mounted on
tmpfs 10M 2.8M 7.3M 28% /dev
[root@speedy:/home/jorj]# df -h /dev/hda3
Filesystem Size Used Avail Use% Mounted on
/dev/hda3 63M 45M 19M 72% /boot
[root@speedy:/home/jorj]# df -h /dev/hda4
Filesystem Size Used Avail Use% Mounted on
/dev/hda4 20G 16G 3.8G 81% /home

I have removed a few old logs from /var/log and have found some stuff from 2005 and 2006 in /var/backups.
What can I safely remove to make more space or can I move stuff into /home?

Thank you jorj

[sammil]

Have you ever run apt-get autoclean and apt-get clean? Might just be a lot ol old debs that have been replaced by upgrades.

Edit. My root partition is almost exactly the same size and only 3.7 GB is in use. I usually run autoclean and clean every other month.

[uteck]

You can run "du -hs /var/backups" to find out how much space they are using. I see only 2.3MB and 3.2 MB on a few systems here, so it may not help. You will get more space running 'apt-get clean' as that will remove all downloaded debs. autoclean will keep the latest version and delete the rest.

[jorj]

Thank ou both. I picked-up 1.2 G with apt-get clean and apt-get autoclean. /var/backups size is 4.3 M.
I wonder why I'm using so much space on /dev/hda1?

Thanks again jorj

[fos]

Welcome back Jorj. I'm glad you are feeling better!

fos

[Lavene]

Another place that have a nasty tendency to grow out of proportion is ~./thumbnails which is basically cached images, all safe to delete as they will be regenerated as needed. The problem is that if you don't clear it every now and then it will save a thumbnail of every picture you have ever opened. Which can amount to several Gb...

Tina

[jpaulb]

Another place that have a nasty tendency to grow out of proportion is ~./thumbnails which is basically cached images, all safe to delete as they will be regenerated as needed. The problem is that if you don't clear it every now and then it will save a thumbnail of every picture you have ever opened. Which can amount to several Gb...

Tina

I cleaned 10000+ out of that folder today trying to get a backup to fit on one DVD

[jorj]

I was trying to download a ML-1710 driver from SamsungUSA and my box locked up twice. I had to reboot.
When I checked the size of /dev/hda1 it was back to 7.9 G apt-get clean and autoclean didn't make a dent in it.
What could have happened - things were so nice for a couple of days.
I did an apt-get update but NOT an upgrade because there was too much to add.

Any suggestions other than stay away from the Samsung site.

jorj

[jpaulb]

When I checked the size of /dev/hda1 it was back to 7.9 G apt-get clean and autoclean didn't make a dent in it.

Any suggestions other than stay away from the Samsung site.

jorj

Which partitions reside in /dev/hda1? check /etc/fstab it will tell. For me it is /boot

Sorry didn't read. if / is on hda1 that means you have /usr /var/ /lib, /usr can take up a lot of room if you down load a lot of docs. If you use mysql the data reside in /var/lib/mysql same with log file if they are not being purged

[jorj]

/dev/hda1 contains everything except /boot and /home which are on separate paritions.

Forgive the following, which will be quite wordy.

The bad harddrive boots into the single user mode, when trying to boot multiuser. There are several messages when booting which say there is no room to write files like mmap and GDM and others.

I am not familiar with single user mode, to copy some of the ~/syslog and paste it into a message.

For a digression; My son brought over a box and he used Ghost to make an image of the bad harddrive. He then rewrote the image to a larger capacity drive and enlarged the partitons. He made the 7.9 G partition 12 G. We thought surely there will be enough room. The new disk boots the same way with the same errors and says there is 100% usage of /dev/hda1 at 7.9 G.

I checked the size with fdisk and it says /dev/hda1 is 12 G, but df -h /dev/hda1 still says 7.9 G at 100%. So now I have two bad drives!

I would greatly appreciate any help. Naturally, I'm mentally kicking myself for not backing-up, but I probably couldn't restore with the drive in this condition.

George

[bhobjj]

I would greatly appreciate any help. Naturally, I'm mentally kicking myself for not backing-up, but I probably couldn't restore with the drive in this condition.

George

You can use the du (disk usage) command to look for large files or directories.
# du -s * | sort -n

If you cannot find any obvious large files to delete such as CD images or videos, you could remove some large files and reinstall after you have everything repaired. Since you don't have enough free space for apt to work, you could manually remove some program files that are not critical for system use.

eg: /usr/lib/openoffice is using 194MB on my system
I could do this (carefully):
# rm -rf /usr/lib/openoffice






For a digression; My son brought over a box and he used Ghost to make an image of the bad harddrive. He then rewrote the image to a larger capacity drive and enlarged the partitons. He made the 7.9 G partition 12 G. We thought surely there will be enough room. The new disk boots the same way with the same errors and says there is 100% usage of /dev/hda1 at 7.9 G.


I prefer the GParted-Clonezilla LiveCD (http://gparted.sourceforge.net/livecd.php). Gparted might work for repairing the partition table so that you actually have 12GB of space.

[uteck]

/var/log can fill up with log messages really fast. This sounds like one of thous times. Use 'du- hs /var/log' to see if that is the problem. If not, it maybe one of the other directories in /var or possibly /tmp.

[jorj]

I think I have a bad file rather then too many files. I say this because everything was working fine with 6.7 Gb in / but then I tried to load the SamsungUSA driver and the box locked (no mouse or keyboard). I had to push the reset and reboot and then I started to get the "no room error message with 7.9 Gb used in / and no room for GDM,mmap, or several other files. We have removed well over 2 Gb (wiping out my win4lin). The drive will still not boot except to put me in single user mode. I guess I'll just have to reinstall Libranet 3.0.
Fortunately, I have win4lin on the Debian drive and can recreate my tax info from raw data in a paper folder. Somewhere on the bad drive are the three Qnn.* files from Quicken that I put in a backup directory. One is too large for a floppy, so I'll try to tar it.
I don't know if I can do a cdrecord in single usermode?

Sorry for sharing all my misery with the forum,but I'm grasping at straws to get this resolved.

jorj aka George

[patrick]

The new disk boots the same way with the same errors and says there is 100% usage of /dev/hda1 at 7.9 G.

Maybe I've missed something, but have you ruled out your box, (BIOS, etc.) rather than the drive(s) as being the problem? Can you mount your drives on another machine to salvage data?

Pat

[jorj]

To thee best of my knowledge, the BIOS is O.K. I use insertable harddrives and my Debian drive seems O.K. I didn't check my Windows XP because I'm using that drive as a temporary repository for the image file made from the from the bad drive.
I use TaxAct in win4lin, but I can't identify my return by the file name, so as I said, it's just a matter of filling out the forms again, but I have all the data.
Sorry for my verbose answers, but I just get carried away.

jorj aka George

[jpaulb]

I checked the size with fdisk and it says /dev/hda1 is 12 G, but df -h /dev/hda1 still says 7.9 G at 100%. So now I have two bad drives!

George

try e2fsck -p if you have an ext2 or ext3 file system. it could be that a super block has false info, only guessing there.

[bhobjj]

I don't know if I can do a cdrecord in single usermode?

jorj aka George

It should work OK. It is worth trying.

Link to a Cdrecord How-To I last edited about 2.5 yrs ago:
http://forums.debian.net/viewtopic.php?t=161

[bluesdog]

Sometimes the disk geometry gets screwed up, but the disk itself is ok

I've used testdisk a few times to restore the proper disk settings, so it may be worth a try for your situation

bhobjj has contributed an excellent how-to (http://www.linuxagora.com/vbforum/showthread.php?p=3544#post3544) regarding file recovery, using photorec, which is part of the testdisk toolset

Testdisk is available from the debian repositories, and is included on many live cd distros