Huge Number of PDL Errors

[John Zalesak]

We are sending messages off site through a secure LAN to LAN tunnel.

We have single IP set up with three different ports.

On two of the ports, we generated about 6,000,000 errors per port in a little over an hour  😯

Here is the error message.

[pdl :PDL :ERR /0:  to_inpc_lab:04/20/2009 14:01:54] read returned error 238 (Connection timed out)

It made the log files Huge.  Good thing it only lasted an hour.

We have seen these errors before but never so many for so fast.

If this had happened over night, it would have shut down the system.

Is there a setting in the engine that controls how quickly the PDL retries on a Connection time out???

Is there any other way to control this so it does not shut down the system???

Topic

[Russ Ross]

One of the ways we protect our system from shut down due to a run away log file is to go into the NetConfig for every site and set a max size for the process log file before Cloverleaf will automatically recycle it.

Here are the steps to follow:

NetConfig –> Process –> Configure –> Automatic Log Cycling –> Threshold in Kilobytes

Russ Ross
RussRoss318@gmail.com

[Max Drown (Infor)]

We’ve had this problem as well with a VPN connection. So far, HealthVision support has not returned a solution or fix to us. We took the same approach as mentioned above and limited the size of the log file.

-- Max Drown (Infor)

[John Zalesak]

I just sent HealthVision an email letting them know we have this issue also.

Thanks for the advice on rolling the log file on size.  I will put that in place also.

If/when we learn something, I will post it here.

Regards

[Chris Williams]

If you are getting time-out errors on these connections, you should check the time-out settings on each piece of hardware in the route. It may not be the source or destination systems, but rather a router or switch somewhere in the middle that has an inappropriate setting.

We had this problem several years ago (on HP-UX) and resolved it by adjusting the tcp_keep_alive_interval from 2 hrs down to 15 minutes. Our culprit turned out to be a switch with a time-out of less than 2 hrs.

[Max Drown (Infor)]

Max Drown wrote:

We’ve had this problem as well with a VPN connection. So far, HealthVision support has not returned a solution or fix to us. We took the same approach as mentioned above and limited the size of the log file.

Support is actively working with me on it, we just haven’t determine a root cause and a fix to my knowledge.

-- Max Drown (Infor)

[John Zalesak]

Max Drown wrote:

Max Drown wrote:

We’ve had this problem as well with a VPN connection. So far, HealthVision support has not returned a solution or fix to us. We took the same approach as mentioned above and limited the size of the log file.

Support is actively working with me on it, we just haven’t determine a root cause and a fix to my knowledge.

If you need another data point for the same problem.  Let me know and I will get you the info.

[Author]

Replies