Steven Lembark wrote:
I have four FAH jobs running on my compute server. I
can "kill -TERM fah6" in about 0.70 sec here, they
start up again and just keep going. FAH is pretty
robust when it comes to restarts; again if you crash
the proc's then it won't be any worse than the outcome
of loosing power: FAH will have to pick up its pieces
and keep going. At least with "halt -f" you'll get
the kernel space cleaned up.
Halt will stop the O/S (see note from manpage, below).
In this case a 'halt -f' would get the system down
about as quickly as possible without just hitting
the reset button.
NOTES
Under older sysvinit releases , reboot and halt should never be
called
directly. From release 2.74 on halt and reboot invoke
shutdown(8) if
the system is not in runlevel 0 or 6. This means that if halt or
reboot
cannot find out the current runlevel (for example, when
/var/run/utmp
hasn't been initialized correctly) shutdown will be called, which
might
not be what you want. Use the -f flag if you want to do a hard
halt or
reboot.
I see your point on stopping FAH. Here, when I do a regular stop, it
has a 17 second wait, can be 60 seconds depending on what it is doing at
the time. That is if it is called by /etc/init.d/zzfah stop. I didn't
have time to type in a lot of commands at the point my P/S was stinking
my room up. You are also correct that FAH is very robust. It writes
its restart point every 3 minutes on this rig so the most it will loose
is about 3 minutes. I have only lost data with FAH once.
I did test the halt -f command last night. I must say, it was fast. It
was literally a few seconds, very few. I did have one file system that
was . . . well . . . a little upset. I use reiserfs and after a fsck,
everything was fine. I also learned to add the -p option to that
command. The halt -f command but did not power off my system.
I learned a lot with this ordeal. One thing is that the P/S's
protection circuit must have worked very well. My mobo is doing just
fine so no damage outside of the P/S itself. I also learned that the
halt -f -p command should be really fast if this happens again.
Keep those thoughts coming.
Dale
:-) :-)
--
gentoo-user@lists.gentoo.org mailing list