On Thu, 28 Oct 2004 14:58:07 +0800 "Tommy Ng" <[EMAIL PROTECTED]> wrote:
> This problem occurred twice in my production system. (v1.3.1 on Solaris > 8) > > One of the 4 CIMD2 links would suddenly stop sending and receiving > messages based on bearerbox access log. Then after a lapse of 6-8 hours > the bearerbox says "SMSC is not alive" and restarts the connection. In > the next second about 2000 requests come flooding through this CIMD2 > connection and causes the system to run out of file descriptors and I > get "System error 24: Too many open files" from smsbox. Kannel stops > functioning after that. > Try increase the rlim_fd_max and rlim_fd_cur in /etc/system You can monitore opened fd count with lsof command also. > After this occurred twice I have now set maximum-queue-length in the > core group configuration but still am not sure why this happens and > whether it will occur again. The SMSC guys say that there were no errors > and the SMSC did not flood the kannel. The 2 failures occurred on 2 > different CIMD2 links to 2 different SMSCs. > > Any ideas? > > Thanks, > Tommy > > -- Best regards, Yury Mikhienko. IT ERP group head, ZAO "Mobikom-Kavkaz" tel: (8632) 704188
