On Friday, 06/22/2007 at 11:50 ZW3, Carlos Bodra <[EMAIL PROTECTED]> wrote: : > DTCUTI002E TCP/IP was very low on TCP connections. Of 285 blocks, 41 are free > after 85 more allocated. > > These messages occur with differences of some days. In TCPIP PROFILE file we > have only SMALLDATABUFFERPOOLSIZE 150 included. All others buffers are default > values. > > What other buffer(s) I should modify (define) to solve this problem. Our
> network has about 300 to 350 at peak moment. The information Miguel gave you will let you get rid of these messages. However, they are really just informational/warnings (I don't know why they are "E" messages) as TCP/IP automatically expands the buffer pool to accomodate needs. The message indicates is that you are getting more TCP/IP activity than you usually do. If you have increased the workload and then you might reasonably expect additional resource consumption in the stack and increasing the initial buffer allocation is justified. If the activity isn't expected, then it may be a sign of trouble or it may be an anomaly and you should ignore it. If you over-allocate memory to a particular buffer pool, you are reducing the amount of memory available for other uses in the stack, so don't get *too* carried away! Look also at the NETSTAT POOL command. Alan Altmark z/VM Development IBM Endicott
