Re: Panic Dumps

• Date: Mon, 28 Aug 1995 15:21:17 PDT
• From:
  "Walter A. Aprile" <aprile@nl.cs.cmu.edu>
• Cc: MOO Cows <submit-mail-moo-cows@ratatosk.yggdrasil.com>
• Content-Length: 1118
• Content-Transfer-Encoding: 7bit
• Content-Type: text/plain; charset=US-ASCII
• In-Reply-To: <199508282135.WAA00823@hinge.mistral.co.uk> from "bartmoss@mistral.co.uk" at Aug 28, 95 02:35:32 pm

bartmoss@mistral.co.uk scribbles :
#
#On my MOOs check points this error is reported.
#
#DUMPIN on safe.db.#1#
#***PANIC: memory allocation (size 16) failed!
#PANIC-DUMPING on ...db.PANIC
#***PANIC              blar
#***REECURSIVE-PANIC   blar
#
#This is basically running out of hard disk space right?

I am not sure about this. Under AIX, if there is not enough disk
space, it simply happens that the dump attempt fails and the 
forked process that tries to dump, simply erases the partly-dumped
database and starts dumping again. 
I have seen this go on and on until

- either there is enough disk space to complete the dump

- or the machine crashes under the doubled load of moo + forked moo

#If so, how under Linux do you find out how much disk space is used, left etc?

Most UN*X's use df

#Any recommandations on how much space should be reserved for the checkpoints?

I'd reserve a space equivalent to the size of your current database 
plus 20 %, just for safety.

BTW, I don't know about you guys, but when my MOO forks and starts
dumping its 25 Megs to disk I always cross my fingers ...


Omaggi, Walter

• Panic Dumps
• From: bartmoss@mistral.co.uk

• Prev: Panic Dumps
• Next: Re: Panic Dumps
• Index: MOO-cows
• Thread: MOO-cows