dangerous situation with shutdown process

[Chuck Swiger]

[ ...crossposting trimmed... ]

Sergey N. Voronkov wrote:
> On Thu, Jul 14, 2005 at 04:17:06PM -0400, asym wrote:
[ ... ]
>> The funny thing about all the replies here.. is that this guy is not saying 
>> that sync doesn't work.
>>
>> He's saying that the timeout built into shutdown causes it to *terminate* 
>> the sync forcibly before it's done, and then reboot.

Good observations.

>> All finger pointing about IDE, SCSI, softupdates, and journals aside.. I 
>> think all he wants/needs is a way to increase that timer.

Sort of.  If you queue up a huge I/O operation which takes several minutes to 
complete, and then also insist upon shutting the machine down before the 
operation completes, well, one of them is going to have to preempt the other.

Extending the shutdown indefinitely is not a reasonable option if the 
two-minute shutdown warning is coming from a UPS which is about to run out of 
battery, and the end result of killing a process that won't politely die when 
SIGTERM'ed to end up with clean filesystems is a better choice.

> If you can't increase shutdown timeout, decrease softupdates timers.
> 
> # tail -3 /etc/sysctl.conf
> kern.metadelay=14
> kern.dirdelay=15
> kern.filedelay=17
> 
> That was my solution for shutdown wait timeout.

This may help, too, but it may not always be enough.

-- 
-Chuck