Weird application freeze problem
list Rich Smrcina
Posting for a friend.... I've been testing out Hobbit in a SLES 10 virtual machine on z/VM. It is a monitoring application based off of Big Brother. So far, it works great except for one weird thing. We have a cron task that runs once a night that does ntpdate to sync the time with an NTP server. If the time is in sync, no problem. However, if the time is out of sync and is adjusted, Hobbit freezes. The tasks still exist, but they just stop doing anything. Apache continues to display the same web page without an update. Stopping the Hobbit daemon doesn't help, in fact, it does nothing. The tasks never stop. I have to recycle the entire system in order to free it up. Has anyone seen a problem similar to this? Any ideas? -- Rich Smrcina VM Assist, Inc. Phone: XXX-XXX-XXXX Ans Service: XXX-XXX-XXXX user-61add9955ef9@xymon.invalid http://www.linkedin.com/in/richsmrcina Catch the WAVV! http://www.wavv.org WAVV 2009 - Orlando, FL - May 15-19, 2009
list Ralph Mitchell
Does your ntpdate have the -B option?? If so, it can make the adjustment
slowly.
From the Gentoo Linux manpage:
-B Force the time to always be slewed using the adjtime()
system
call, even if the measured offset is greater than +-128 ms.
The
default is to step the time using settimeofday() if the
offset
is greater than +-128 ms. Note that, if the offset is
much
greater than +-128 ms in this case, that it can take a long
time
(hours) to slew the clock to the correct value. During
this
time. the host should not be used to synchronize clients.
From the NOTES section of the adjtime(3) man page:
The adjustment that adjtime() makes to the clock is carried out in
such
a manner that the clock is always monotonically increasing. Using
adj-
time() to adjust the time prevents the problems that can be caused
for
certain applications (e.g., make(1)) by abrupt positive or
negative
jumps in the system time.
Ralph Mitchell
▸
On Mon, Sep 8, 2008 at 10:10 PM, Rich Smrcina <user-cf452ff334e0@xymon.invalid> wrote:
Posting for a friend.... I've been testing out Hobbit in a SLES 10 virtual machine on z/VM. It is a monitoring application based off of Big Brother. So far, it works great except for one weird thing. We have a cron task that runs once a night that does ntpdate to sync the time with an NTP server. If the time is in sync, no problem. However, if the time is out of sync and is adjusted, Hobbit freezes. The tasks still exist, but they just stop doing anything. Apache continues to display the same web page without an update. Stopping the Hobbit daemon doesn't help, in fact, it does nothing. The tasks never stop. I have to recycle the entire system in order to free it up. Has anyone seen a problem similar to this? Any ideas? -- Rich Smrcina VM Assist, Inc. Phone: XXX-XXX-XXXX Ans Service: XXX-XXX-XXXX user-61add9955ef9@xymon.invalid http://www.linkedin.com/in/richsmrcina Catch the WAVV! http://www.wavv.org WAVV 2009 - Orlando, FL - May 15-19, 2009
list Rich Smrcina
I'll pass this on, thanks.
▸
Ralph Mitchell wrote:Does your ntpdate have the -B option?? If so, it can make the adjustment slowly.
From the Gentoo Linux manpage:
-B Force the time to always be slewed using the adjtime() system
call, even if the measured offset is greater than +-128 ms. The
default is to step the time using settimeofday() if the offset
is greater than +-128 ms. Note that, if the offset is much
greater than +-128 ms in this case, that it can take a long time
(hours) to slew the clock to the correct value. During this
time. the host should not be used to synchronize clients.
From the NOTES section of the adjtime(3) man page:
The adjustment that adjtime() makes to the clock is carried out in such
a manner that the clock is always monotonically increasing. Using adj-
time() to adjust the time prevents the problems that can be caused for
certain applications (e.g., make(1)) by abrupt positive or negative
jumps in the system time.
Ralph Mitchell-- Rich Smrcina VM Assist, Inc. Phone: XXX-XXX-XXXX Ans Service: XXX-XXX-XXXX user-61add9955ef9@xymon.invalid http://www.linkedin.com/in/richsmrcina Catch the WAVV! http://www.wavv.org WAVV 2009 - Orlando, FL - May 15-19, 2009
list Henrik Størner
▸
In <user-6a01c5c4e516@xymon.invalid> Rich Smrcina <user-cf452ff334e0@xymon.invalid> writes:
I've been testing out Hobbit in a SLES 10 virtual machine on z/VM. It is a monitoring application based off of Big Brother. So far, it works great except for one weird thing. We have a cron task that runs once a night that does ntpdate to sync the time with an NTP server. If the time is in sync, no problem. However, if the time is out of sync and is adjusted, Hobbit freezes. The tasks still exist, but they just stop doing anything. Apache continues to display the same web page without an update. Stopping the Hobbit daemon doesn't help, in fact, it does nothing. The tasks never stop. I have to recycle the entire system in order to free it up.
Generally, Hobbit gets pretty confused if time goes back. If you must
play around with the system clock regularly, I'd suggest stopping
Hobbit while doing so.
It will wreak havoc on your statistics, though ..
Regards,
Henrik