TLI connection problems with Online 7.12.UC1 -Reply

Bill I am currently working through a similar problem with 7.12.uc1 which appears related to the setting of SHMTOTAL. This was 0. I changed it to 128Meg. (last friday) Symtoms similar to those you report occurred. In my case shared memory connections were still possible, it was like the port had become blocked. I reset SHMTOTAL to 0 (since that was all that had changed) and the problems [seem to] have gone away. ( We are still taking a watching brief on this and the instance in running with AFDEBUG set in case it recurrs.) On the same machine I have another instance with SHMTOTAL of 250Meg with no problems. The box has 1 GIG of Ram. HTH -- -- Peter Tashkoff NZ Kiwifruit Project Consultant/DBA Marketing Board. Standard Disclaimers apply This posting may not be used by any commercial organisation to vilify another. >>> Spencer, Bill <SPENCERB@angec1.agw.bt.co.uk> 26/July/1996 02:22am >>> Hi there, We are running Online 7.12.UC1 on a Pyramid Nile and have encountered a problem with one of our engines. We have been hit twice with error -27001 - Read error occurred during connection attempt. The server has effectively been lost to all existing connections. The protocol used is TLITCP using local loopback (clients on same machine). The network connections were checked out and were fine. Bouncing the engine brought everyone back online. Has anyone has this problem before, any info would be appreciated. Mail address spencerb@angec1.agw.bt.co.uk