Intralink 10.0 Error: "Unable to extract a server stream from the connection"

[ScottPearson]

'Chillers,

Has anyone encountered the followingerror message in the Event Manager?

"Unable to extract a server stream from the connection"

A quick search of this Forum and PTC's Knowledge Base did not uncover theexact message.

I have a soleuser who is having intermittent network connection issues during his ProE / Intralink 10.0 sessions. He seems to be able to Check-In most files one-at-a-time, but when attempting to Check-In an assembly with multiple parts he is consistently getting the error as the multi-object Check-In fails.

All signs, of course, point to a shakey network connection but his system is new and he doesn't have any other network-connectivity symptoms beyond this Intralink 10.0 issue. This just started today. No major system changes made today (that I know of).

I'm having IT troubleshoot/verify the integrity of the user's network drop but I wondered if any of you have encountered this error message and if so what did you find was the ultimate culprit?

WF4 m210 / Ilink 10.0 m030

Win 7 / 64-bit / Lenovo laptop

Other users on similar (identical) systems don't have this problem.

I have already wiped out and recreated his local cache: problem still occurring.

TIA.

[TomU]

We have the issue occasionally here on a couple of computers. We have not been able to determine the root cause. Re-attempting the check-in usually solves the problem. I think it may be related to the anti-virus software, but haven't been able to prove it. We are using WF5 M120 with Windchill 9.1 M050.

Tom U.

[wpalm_99]

Tom/Scott,

You could try increasing your KeepAliveTimeout value for Apache. Depending
on the version you're using, the default value may be different, so I won't
give you a value to use. I've seen this cause huge performance issues for
remote users when a customer set this to 30 based on a blind recommendation
as it turned out they were actually lowering the value that somebody had
already changed.

Look for the setting in the <apache>\conf directory. Which .conf file it's
in also depends on the version. You also need to be careful if a value is
assigned in multiple .conf files. The last one read by Apache will be the
one that's used. For example, if it's set in httpd.conf, but also set in
additions.conf, the value from httpd.conf will only be used if it occurs
after the Include statement for additions.conf. In other words, there's a
heirarchy, but it's based on the order in which the files are read starting
with httpd.conf.

Once you find the value that's actually used, double it and restart Apache.
My guess is that you're at the default value of 15. If so, you can
increase 30 with no problem. If you are going higher than that, you'll
probably also need to double the value for MaxKeepAliveRequests (default
100).

If you do find that the property occurs in multiple files and you're not
sure which one is being used, reply and attach the files and I'll let you
know what to do from there.


Regards,
Bill

[TomU]

PTC tech support already had us increase this to 60 with no change whatsoever. Thanks for the info though.

Tom U.

[wpalm_99]

Tom,

Did you also change the MaxKeepAliveRequests to 200?


Regards,
Bill

[BrianToussaint]

I have had some of the same problems with some users being really slow. I've put it off to the new cable that was run for a new area. There are some other areas that are extremely slow also.

I looked for the two settings. The only one I can find is the KeepAliveTimeout and it is in the load point config.xml and config.properties. I cannot find the MaxKeepAliveRequests anywhere.

I did see that the KeepAliveTimeout is set to 15.

Is this different for Apache inside of Windchill 10.1? How/where should I change mine? Is there any guide for tuning Apache?

Thanks,

Brian Toussaint
CAD Administrator

Hoshizaki America, Inc.
"A Superior Degree Of Reliability"
618 Hwy. 74 S., Peachtree City, GA 30269