inline
Dirk.
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Servers Alive
Sent: Thursday, March 17, 2005 10:28 PM
To: [email protected]
Subject: [SA-list] SA Service Unexpectedly Terminating
I am running SA as a service on Windows Server
2003.
Lately I have been experiencing the service
terminating unexpectedly. I have debug logging set to maximum, and every
time SA crashes, it's right in the middle of writing the log file. A log
file sample from such an instance looks like this:
Monday, March 14, 2005 6:54:06 AM URL check
contains string: Anoka-Hennepin(CON-14)
Monday, March 14, 2005 6:54:06 AM URL check took 391 ms
Monday, March 14, 200Monday, March 14, 2005 6:54:33 AM Servers Alive version 5.0.1680.3
Monday, March 14, 2005 6:54:33 AM Running on Microsoft Windows Server 2003 Family Standard Edition (3790)
Monday, March 14, 2005 6:54:06 AM URL check took 391 ms
Monday, March 14, 200Monday, March 14, 2005 6:54:33 AM Servers Alive version 5.0.1680.3
Monday, March 14, 2005 6:54:33 AM Running on Microsoft Windows Server 2003 Family Standard Edition (3790)
or:
Sunday, March 13, 2005 5:47:06 AM URL check
contains string: Anoka-Hennepin(CON-14)
Sunday, March 13, 2005 5:47:06 AM URL check took 344 ms
Sunday, March 13, 2005 Sunday, March 13, 2005 5:47:20 AM Servers Alive version 5.0.1680.3
Sunday, March 13, 2005 5:47:20 AM Running on Microsoft Windows Server 2003 Family Standard Edition (3790)
Sunday, March 13, 2005 5:47:06 AM URL check took 344 ms
Sunday, March 13, 2005 Sunday, March 13, 2005 5:47:20 AM Servers Alive version 5.0.1680.3
Sunday, March 13, 2005 5:47:20 AM Running on Microsoft Windows Server 2003 Family Standard Edition (3790)
As you can see, SA crashes right in the middle of
writing to the log file, and picks up right where it left off when the service
restarts itself. This happens on average two or three times a day.
It's only been a week, but I haven't yet noticed any patterns in the times that
it crashes -- I'll keep you updated though.
First things that come to my brain:
1. Is the log file for each check cycle
written all at once at the end of a cycle? Or is each line written in
real-time *during* the check cycle itself?
<dirk>During
the cycle but the OS caches parts and that's why the last lines aren't always in
tthe file (except if you STOP the checking - were we force the log to be closed
and reopened - and a close of the file seems to tell the OS to flush
everything)
2. All Windows Event Log reports is that the
service terminated unexpectedly. The moon icon remains in the tray, and
disappears when you put your mouse over it (common when a Windows program
crashes).
<dirk>Try the
version on http://www.woodstone.nu/salive/sr.asp
Several bugs are
solved some of them caused a runtime error (only visible if the GUI was visible
ofcourse)
3. I have a second computer set to
authenticate to the server running SA and check the service is running. It
checks every 30 seconds. I don't think this is related to SA crashing
however, because the 30-second checks and the crashes don't coincide (that I can
tell anyway).
<dirk>Don't
think it's linked.
4. SA crashing does not appear to coincide
with any of the server checks failing... the services I check have all gone up
and down many times in the past week, and 99.9% of the time, SA does all the
checks and alerts flawlessly. .01% of the time, it seems to crash on
me.
Any thoughts from anybody on this? I went
into detail in hopes that I'd trigger something in somebody's brain out there
regarding this problem. Thanks.
- Eric
