81002 Error

lbegley · Post by **lbegley** » Fri Jan 21, 2005 8:18 am

I have looked thru both DSXchange and DevNet (and also have a case open with Ascential) for an issue regarding 81002 (timeout).

This is the message we get:

Failed to connect to host:FSS00DS2, project: UV
(the connection is broken (81002))

We receive this error approximately every 24 hours.
We can boot the server and all is OK for the next 24 hours.

We set the inavtivity timeout to Do Not Timeout and this has 'resolved' this issue.

I would thing that there should be some value to using the timeout.

Sainath.Srinivasan · Post by **Sainath.Srinivasan** » Fri Jan 21, 2005 8:22 am

It is used to avoid any process left open overnight to have the lock beyond certain time. i.e. if anybody leave a job opened, the job will be killed after that duration.

Having said that, the DataStage process itself should not be affected by that.

Secondly, it has been observed that the hanging zombie process such left over ate more memory than the connection itself.

Luk · Post by **Luk** » Fri Jan 21, 2005 11:00 am

if you use windows 2003 it may be sth like this:

http://www.mail-archive.com/u2-users@li ... 05693.html

>This seems to be common problem on windows 2003. One of the ?>resolutions
>is
> to go to services menu under administration tools. Select the properties
> for universe RPC and under the login tab select to allow this service to
> interact with the desktop.
>
> Under windows 2003 there are a number of security features and I suspect
> that part of this is to close down unrecognised sockets that have not been
> active for a while. I believe there is somewhere where you can keep this
> active, but have not had time to explore.

lbegley · Post by **lbegley** » Fri Jan 21, 2005 11:04 am

Thanks for the reply.

We are Windows 2000 for our servers.
I had seen this (changes service to interact with desktop) but problem still existed after making the change.

winterb1 · Post by **winterb1** » Tue Jan 25, 2005 10:20 am

We have the same environment (DataStage 7.x, Win2k), and have run into the same problem. A reboot has fixed the problem. Has anyone found a good solution to this problem?

--Brian

ogmios · Post by **ogmios** » Tue Jan 25, 2005 11:06 am

Log in via DataStage Administrator and either deactivate the timeout, or put it to a much higher value.

Ogmios

ray.wurlod · Post by **ray.wurlod** » Tue Jan 25, 2005 4:59 pm

This problem can occur if defunct processes continue to hold sockets open (even though they're in a state of waiting for timeout).

In some cases, the held sockets prevent the RPC service from starting.

This is why re-booting made the problem go away - removing the operating system also removes any sockets!