When this is happening, check for open transactions on the SQL server or
requests which are currently being processed. Also set your querytimeout
as low as you can tolerate and turn on logging, level 2 or 3 if you can
manage, 1 otherwise. Sometimes I find that code can inadvertently create
a 'select * from table' statement and timeout the server just because
the sql server takes so long to respond. You might also want to check
your SQL server's indexing. Lastly, it might not have anything to do
with SQL, and you should check your other externals, such as email and
COMs, as these can sometimes tie up the server as well. One other
thought, if you are pretty sure this will occur again, set a
once-a-minute cron to record the current and max thread counts to a log
(file or db). This should be able to show you if your problem is
instant, of if your load builds up to the stopping point. If the second
is the case, you might be able to warn yourself and gather more
information. You might also want to save db connection count and memory
heap size. In almost all cases that I've seen, these problems are caused
by an external, and Witango simply runs out of threads or time to get
past it. If Witango is the cause, I can usually see this in the logs. Is
the CPU pinned when service is stuck? Does it eventually drop? Does
witangoevents.log contain any information at all? Is the IIS plugin 065
as well?

One thing to consider, 5.5 is due to be out soon, and could very well
solve this for you, more so because it can better manage db connections
and other externals, if you haven't already, you might want to start
testing the beta.

Robert

-----Original Message-----
From: Dave Machin [mailto:[EMAIL PROTECTED] 
Sent: Tuesday, August 03, 2004 11:24 AM
To: [EMAIL PROTECTED]
Subject: Witango-Talk: WiTango .65 unstable

We've started having some troubling stability issues with WiTango .65 on
W2K SP4.

Pages stop loading and it seems that WiTango is not responding.
Connections to IIS keep stacking up until we get about 260 connections
and eventually the page requests time out.  IIS logs write code 500
errors but the WiTango log is completely missing any entries for the
time until I kill and restart the process.  The log entries prior to the
missing section seem completely normal.  There are no errors in the
WiTango log and no events in the witangoevents.log file.  WiTango does
not respond when asked to stop from Services and I have to kill it using
an application in order to restart it - at which point everything seems
fine for several hours or days.

The server processor is not overly busy, but it isn't completely idle
either.  During the outage the databases the applications talk to are
free of any blocks, so it doesn't seem that WiTango is waiting for
database response.  The database is MS-SQL 2000, and the ODBC driver is
version 2000.85.1022.00.  

I'm at a bit of a loss as to what to even look at as to the possible
cause of the problem.  Any idea what I should check?

Dave Machin 


E-Mail. [EMAIL PROTECTED]
Tel.  805.614.0123 x 30
Address: 3201 Airpark Drive, Suite 104
Santa Maria, CA
93455________________________________________________TOUNUBCRBE G t
htp//wwwiano.omdeelpe/milis.t

________________________________________________________________________
TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf

Reply via email to