🚀 go-pugleaf

In article <1992Sep11.093740.288@gamelan> thf@zelator.in-berlin.de (Thomas Funke) writes:
>
>Setting a highwater mark for the swapfile is'nt very recommendable: It will  
>crash (stop) the NeXT if the mark is reached. How do other UNIX-systems  
>handle this ? AIX starts killing processes, if there is not enough  
>paging-space. But still the system is usable. 
>

Here's my horror story:
My machine locked up because the swapfile filled up the available space,
so I rebooted, only to discover that for some reason the swapfile had *not*
been removed, and the machine would not reboot.  Being a unix novice, I
spent several hours figuring out and fixing the whole mess.  First, I found
that the reboot had not removed the old swapfile as expected.  No problem,
I'll just remove it.  Unfortunately, at the point in the boot process that
I had access to do such things the file system was mounted as read-only.
It tookquite a while to figure that out, then another while to discover
the remount command (i think that was it), finally remove the swapfile,
and reboot successfully.  This whole process was hampered by the fact that,
since the filesystem was completely full, I was unable to read any man
page that had not been previously read (i.e., formatted).  Luckily I had
read the man page for mount, or no telling how long it would have taken
to figure out the whole deal.  Anyway, after this happened I set a highwater
mark and I was *happy* that the next time the machine "only" locked up.

Lusty Wench