No Space on HD

[brandondaly]

I have a server with Sco-Unix 5.04 installed. A couple of days ago the server started to display the error message

NOTICE:HTFS: No space on dev hd

I cannot now get into the system as it does not bring up a logon prompt. Does any body know the likely cause of the problem and a possible solution?

 

[edfair]

Time for the EBD and fsck.
Then get it mounted and clean out some space.
One of the jobs of the administrator is to monitor the filesystems to insure you have free space. Ed Fair
unixstuff@juno.com
Any advice I give is my best judgement based on my interpretation of the facts you supply. Help increase my knowledge by providing some feedback, good or bad, on any advice I have given.

 

[brandondaly]

I managed to get a logon prompt after a couple of hours, I cleared out the log files, but I still low on disk space, the major problem is I don't know whats been filling it up, the system was taken out of general use 6 months ago and is now only used for reference purposes.
Is there a way to list, say, the top ten sized files?

 

[pcunix]

See

http://pcunix.com/Unixart/nospace.html

for a more complete treatment of what to do. Tony Lawrence
SCO Unix/Linux Resources

http://pcunix.com

tony@pcunix.com

 

[jcsalfer]

I've found by using the find command I can quickly locate large files causing problems. Using a +1000000c will locate all files larger than 1MB which shouldn't be many on the root filesystem.

find . -mount -size +1000000c -print | more

I also use the du -k option quite often to find out how many K are in a given directory and subdirectories.

Good Luck,
John.Salfer
Voyageur Software

 

[romoon66]

I've found that if your system is barking about space, that it is more often then most a sudden occurence. So I would suggest looking for larger files first then work your way down. I like using the following to find files larger then 5mb

find / -size +10000

This will list the files larger then 5mb. The system default is 512k per per n listed, 2000 (1mb) and so on.

Happy Hunting.

 

[Guest_imported]

We've had a simular problem from time to time. Initially we would since getting a root login and then trying to work between the no space errors is very difficult, just do a hard reset and when the system came back up the problem disappeared leaving us puzzled. Software support suggested it was a runaway print spool issue. What we found was a cron job generates a file error (open/read) and respawning nature of cron results in an endless loop effect. So we will find a huge crout file (all available allocated space) in /tmp. Only the first few lines are meaningful (the actual error cause, which the code lacked any method of handling, a problem you'll want to address) the rest is just the error repeated.

 

[KarveR]

Other things to look at:
/usr/adm/syslog
/usr/adm/messages
your mail queue (/usr/spool/mqueue or /usr/spool/mail)
/tmp ***************************************
Party on, dudes!

 

[trikk]

One little tidbit..

the command dfspace works well on SCO UNIX...gives a more easily understood respesentation of disk usage

 

[KarveR]

dfspace works well under SCO 5.0.5 unless the server has its head up its .... anyway, my point is (after recent experience, dont trust dfspace except after you've just rebooted the server. ***************************************
Party on, dudes!