11/07/2012 6:25 AM
post96939
|
Hi Sebastián,
Nobody can help you without additional info. What is your network structure? What output of following commands?
sin ar
sin ir
sin ver
sin ti
But it's better try to upgrade from TCP/IP 4.25 to TCP/IP 5.10 and use Tcpip manager instead of Socket.
Also we don't know what happens in your network, how applications use the network, etc. Try to localize the problem. Can
you reproduce this issue with minimal networking configuration with minimum of apps?
Respectfully,
Oleg
07.11.2012, в 14:52:04, Sebastián Cerella написал:
> Hi, there has been a problem with the "Socket" (QNX 4.25) to a disturbance in the management network of the company .
"Socket" consumes memory and CPU disabling the node (image in attachment ). It has happened in many servers at the same
time but not in others connected to the same network. Versions of "Socket" used in these servers with the problem is 4.
25K and 4.25H. Versions of "Proc" used in these servers is 4.25Q. The "resolv.conf" is configured with the nameservers.
> Please, somebody I will guide to fix it?
> Thank you very much.
> Sebastian Cerella
>
>
>
> _______________________________________________
>
> General
> http://community.qnx.com/sf/go/post96935
> To cancel your subscription to this discussion, please e-mail general-qnx4-unsubscribe@community.qnx.com
> <Consu.jpg>
|
|
|
01/14/2013 2:08 PM
post98551
|
I have exactly the same problem. I have 3 servers running Socket and all of them periodically run into high CPU
utilization with Socket along with the software clock running slow (losing up to 10 minutes per hour). The other 25
nodes on the network that don't run Socket are not affected. I am running Proc 4.25Q and Socket 4.25H. When it occurs,
all 3 simultaneously experience the problem. I am sure it is triggered by some sort of network traffic.
I have not found a solution to this problem. I don't know if upgrading to TCPIP 5 is an option for me.
|
|
|