Oninit cdi / Archive
Talk to an expert →

25580 error problem

Posted in 1999

Topics: Performance & Tuning, Connectivity: ESQL/C, 4GL & Embedded SQL, Networking & sqlhosts Configuration, Platform-Specific Issues, Versions, Editions & End-of-Life

Adam Williams — — source: Informix-list mailing list archive (1991-1998)

I have Informix IDS 7.30.UC7-1 on a Redhat Linux 6.1 box (stock 2.2.12-20smp kernel), and 4gl 7.20.UD6 on an IBM RS/6000 running AIX 4.2.1.0. I need to run the 4gl apps on the RS/6000 and use the engince on the RedHat box ( much more powerful machine). I have IDS 7.24.UC7 on the RS/6000 as well, which we intend to shutdown after the other engine works. Running against the local engine the 4gl applications perform normally. Against the remote engine (on the RedHat box) they work, but occasionally die with a 25580 network error. This seems to only happen on large jobs, and usually takes awhile. There are no messages on the RedHat box concerning an error (checked onstat -m, dmesg, and /var/log/messages). The sql hosts file on the RS/6000 is: MIE onipcshm MIE MIE <--local engine MIENET onsoctcp *mie onlineBARNET onsoctcp barracuda online <--remote engine The tables and indexes in both engines are identical. The sql hosts file on the RedHat box is: BARIPC onipcshm BARIPC BARIPC BARNET onsoctcp *barracuda online I've tried tuning lots of things, but am at wits wnd with this one. I start the IDS engine on the Linux box with: echo "16384" > /proc/sys/fs/file-max echo "65536" > /proc/sys/fs/inode-max echo "5 128 64 256 15 3000 500 1884 2" >/proc/sys/vm/bdflush echo "0" >/proc/sys/vm/overcommit_memory echo "262144000" > /proc/sys/kernel/shmmax $INFORMIXDIR/bin/oninit Local process run on the Linux box with no problems. Any help would be greatly appreciated.

↪ replying to Adam Williams
Reinhard Habichtsberg — — source: Informix-list mailing list archive (1991-1998)

On Mon, 29 Nov 1999 13:23:12 -0500 (EST), Adam Williams <adam@morrison-ind.com> wrote: >(..) > Against the remote engine (on >the RedHat box) they work, but occasionally die with a 25580 network error. >This seems to only happen on large jobs, and usually takes awhile. There are >no messages on the RedHat box concerning an error (checked onstat -m, dmesg, >and /var/log/messages). > Did you ask this some weeks ago? Maybe it is a silly idea, but it seems that the tcp connection is broken due of no traffic. Look for the refering kernel parameters. HTH Reinhard

Related threads

Glossary