On 01/26/2018 04:29 PM, Stephen Morris wrote:
On 25/1/18 9:26 am, Patrick O'Callaghan wrote:
> On Wed, 2018-01-24 at 14:01 -0700, stan wrote:
>> On Wed, 24 Jan 2018 12:03:54 +0000
>> Patrick O'Callaghan <pocallaghan(a)gmail.com> wrote:
>>
>>> On rebooting today my system took an age to shutdown, for no apparent
>>> reason but possibly related to NFS mounts.
>> I run a daemon to feed entropy into the kernel pool. It sleeps most of
>> the time, waking up only when the pool drops below a threshold. If I
>> shutdown while it is sleeping, it takes a minute and a half before
>> systemd sends it a kill -9, unless it wakes up in the interval. If I
>> kill the process before shutdown, there is no wait. Perhaps
>> something like this is happening to you (the NFS mounts?).
> This would imply that the daemon isn't responding to a TERM signal,
> hence the timeout. I thought NFS did obey TERM, but a glance at the man
> pages doesn't show any reference to signals, so the answer must remain
> a mystery.
I get the same issue on my system but for me it is completely random as
to when it happens and when it doesn't. The last time it happened , it
paused for a minute or so after the normal watchdog not stopping
message, and just before it eventually decided to shutdown another
message was displayed that looked like it might have been relative to
mounts, but it disappeared too quickly to be sure.
Have you ever checked the status of the processes (such as via top or
ps with an "-o s" option) to see if they're in a D state? If so,
they're
in an I/O wait state and poking them via signal() won't wake them. They
either timeout or you just pull the rug out via a reboot.
----------------------------------------------------------------------
- Rick Stevens, Systems Engineer, AllDigital ricks(a)alldigital.com -
- AIM/Skype: therps2 ICQ: 22643734 Yahoo: origrps2 -
- -
- IGNORE that man behind the keyboard! -
- - The Wizard of OS -
----------------------------------------------------------------------