I'm seeing the following at one of my customers sites:

Error #1100 A connection to the server at xxx.xxx.xxx.xxx cannot be
established ('Socket error: An existing connection was forcibly closed by
the remote host. 10054, on API 'recv'')


It doesn't happen all of the time, but frequently enough to be a bother.  I
believe that users who are already connected to the server don't have any
problems while this occurs, and after a while the user can then connect to
the server without any problems.

Not positive about this, but it appears that the issue started after
updating the server engine to 2.03b21 as that is when they started
mentioning the issue.

Anyone have any suggestions on where I should look for the cause of this?

--

---------------------------------------
 Terry Swiers
 Millennium Software, Inc.
 http://www.1000years.com
 http://www.atrex.com

Now shipping Atrex 13.  For more information go to
 http://www.atrex.com/news.asp

Atrex Electronic Support Options:
 Atrex Knowledgebase: http://www.atrex.com/atrexkb.asp
 Email: mailto:support@atrex.com
 Newsgroup: news://news.1000years.com/millennium.atrex
 Fax: 1-925-829-1851
 Phone: 1-925-828-5892 (M-F, 9a-5p Pacific)
 ---------------------------------------

Terry,

<< I'm seeing the following at one of my customers sites:

Error #1100 A connection to the server at xxx.xxx.xxx.xxx cannot be
established ('Socket error: An existing connection was forcibly closed by
the remote host. 10054, on API 'recv'')

It doesn't happen all of the time, but frequently enough to be a bother.  I
believe that users who are already connected to the server don't have any
problems while this occurs, and after a while the user can then connect to
the server without any problems.

Not positive about this, but it appears that the issue started after
updating the server engine to 2.03b21 as that is when they started
mentioning the issue.

Anyone have any suggestions on where I should look for the cause of this? >>

The only changes in B21 related to this was this:

http://www.elevatesoft.com/incident?action=viewrep&category=edb&release=2.03&incident=3285

and it shouldn't have any adverse effects upon anything else.  In fact, if
you didn't encounter this bug in 2.03 B20, then you are running the exact
same remote session code.

When are the users seeing this issue ?  Is it during login, during operation
of the application, etc. ?

--
Tim Young
Elevate Software
www.elevatesoft.com

Tim,

> The only changes in B21 related to this was this:
> and it shouldn't have any adverse effects upon anything else.  In fact, if
> you didn't encounter this bug in 2.03 B20, then you are running the exact
> same remote session code.

Then I'm running the same remote session code.  The only change with our
last build was to update to 2.03 B21.


> When are the users seeing this issue ?  Is it during login, during
> operation of the application, etc. ?

It is during initial startup of our app or when switching databases.  I've
not traced the error to deteremine where the error is exactly, but it is
either on the session connection or possibly when opening the table
"location" in our database as the error message has "location: ElevateDB" at
the beginning.   I don't have a session or database named "location", so I
assume it is the table name.

If it is the table being opened, the thing that confuses me about the
message is that by the time the location table is opened, the session has
already been implicitly opened in the code and verified that I have a
connection.

--

---------------------------------------
 Terry Swiers
 Millennium Software, Inc.
 http://www.1000years.com
 http://www.atrex.com

Now shipping Atrex 13.  For more information go to
 http://www.atrex.com/news.asp

Atrex Electronic Support Options:
 Atrex Knowledgebase: http://www.atrex.com/atrexkb.asp
 Email: mailto:support@atrex.com
 Newsgroup: news://news.1000years.com/millennium.atrex
 Fax: 1-925-829-1851
 Phone: 1-925-828-5892 (M-F, 9a-5p Pacific)
 ---------------------------------------

Tim,

Just got off the phone with them and have some new information.

They finally got to the point where no user could log into the system so
they called.  Connected into the server and found that I was unable to
connect to the server at all.  The standard edbsrvr was sitting at about
140MB of memory allocated (as reported from task manager) and 80% cpu usage.
This is with no users connected to the database server.

I couldn't log in with my server manager software either, getting error 1100
just trying to connect.  Stopped the server engine service, restarted, and I
was able to connect back in.

I'm going to do some additional testing on my side to see if I can replicate
the problem.  It almost appears as if the sessions objects are being held
onto rather than being freed even after the users disconnect and ultimately
hits some internal limit which is preventing additional connections.

I'll let you know what my testing shows.

--

---------------------------------------
 Terry Swiers
 Millennium Software, Inc.
 http://www.1000years.com
 http://www.atrex.com

Now shipping Atrex 13.  For more information go to
 http://www.atrex.com/news.asp

Atrex Electronic Support Options:
 Atrex Knowledgebase: http://www.atrex.com/atrexkb.asp
 Email: mailto:support@atrex.com
 Newsgroup: news://news.1000years.com/millennium.atrex
 Fax: 1-925-829-1851
 Phone: 1-925-828-5892 (M-F, 9a-5p Pacific)
 ---------------------------------------

Terry,

<< I couldn't log in with my server manager software either, getting error
1100 just trying to connect.  Stopped the server engine service, restarted,
and I was able to connect back in. >>

Did you check the log to see if there was anything in there ?  It should
indicate why the connections were being rejected.

Thanks,

--
Tim Young
Elevate Software
www.elevatesoft.com

Tim,

> Did you check the log to see if there was anything in there ?  It should
> indicate why the connections were being rejected.

Got a report of the same exact issue at another client.  Unfortunately both
clients are gone for the day so I won't be able to check on this until
tomorrow.

--

---------------------------------------
 Terry Swiers
 Millennium Software, Inc.
 http://www.1000years.com
 http://www.atrex.com

Now shipping Atrex 13.  For more information go to
 http://www.atrex.com/news.asp

Atrex Electronic Support Options:
 Atrex Knowledgebase: http://www.atrex.com/atrexkb.asp
 Email: mailto:support@atrex.com
 Newsgroup: news://news.1000years.com/millennium.atrex
 Fax: 1-925-829-1851
 Phone: 1-925-828-5892 (M-F, 9a-5p Pacific)
 ---------------------------------------

Tim,

> Did you check the log to see if there was anything in there ?  It should
> indicate why the connections were being rejected.

Got access to one of the customers servers. In the hours leading up to the
incident I see aproximately 30 AVs corresponding to a Update Row call, 2 AVs
corresponding to Free Cursor calls, 1 entry for a table does not exist in a
drop table call, and 1 entry indicating that a filter could not be created
because it was malformed.

There are no 1100 entries in the server log at all.

--

---------------------------------------
 Terry Swiers
 Millennium Software, Inc.
 http://www.1000years.com
 http://www.atrex.com
 ---------------------------------------

Terry,

<< Got access to one of the customers servers. In the hours leading up to
the incident I see aproximately 30 AVs corresponding to a Update Row call, 2
AVs corresponding to Free Cursor calls, 1 entry for a table does not exist
in a drop table call, and 1 entry indicating that a filter could not be
created because it was malformed. >>

The AVs are most likely causing a cascading failure.  The AV issues appear
really close to this incident report:

http://www.elevatesoft.com/incident?action=viewrep&category=edb&release=2.03&incident=3247

And none of these issues occurred with 2.03 B20 ?  I just went through the
fixes for B20 again, and none of them were anywhere near this type of
functionality, so I've got to think that this was happening prior to B21
also.

--
Tim Young
Elevate Software
www.elevatesoft.com

Tim,

> The AVs are most likely causing a cascading failure.  The AV issues appear
> really close to this incident report:
>
> http://www.elevatesoft.com/incident?action=viewrep&category=edb&release=2.03&incident=3247
>
> And none of these issues occurred with 2.03 B20 ?  I just went through the
> fixes for B20 again, and none of them were anywhere near this type of
> functionality, so I've got to think that this was happening prior to B21
> also.

Yes.  Those AVs are the exact issue that we've been working on for a while
with the 9999's during Update Row .

And yes, it is quite possible that it was happening prior to B21 but was
simply not being reported to us.


--

---------------------------------------
 Terry Swiers
 Millennium Software, Inc.
 http://www.1000years.com
 http://www.atrex.com

Now shipping Atrex 13.  For more information go to
 http://www.atrex.com/news.asp

Atrex Electronic Support Options:
 Atrex Knowledgebase: http://www.atrex.com/atrexkb.asp
 Email: mailto:support@atrex.com
 Newsgroup: news://news.1000years.com/millennium.atrex
 Fax: 1-925-829-1851
 Phone: 1-925-828-5892 (M-F, 9a-5p Pacific)
 ---------------------------------------

Terry,

<< Yes.  Those AVs are the exact issue that we've been working on for a
while with the 9999's during Update Row .

And yes, it is quite possible that it was happening prior to B21 but was
simply not being reported to us. >>

Okay, we're on the same page then.

--
Tim Young
Elevate Software
www.elevatesoft.com