Remote sqlturbos hang in Informix 5.0

Hi, I need some new ideas to help solve a problem. The informix hotline has not been helpful. It is hard to get past the first line of defense. We have been having this serious problem with sqlturbos that hang, or, as we refer to it - freeze. Our environment: Informix Version 5.02.UC3 UNIX System V/88 Release 4.0 Version R40V4.3 Motorola Processor Our database: Two machines using Informix Online/Star. The data between them is replicated using two phase commits (our own coding in SQL). One machine is basically a mirror of the other. Updates to the database can occur from either machine. Our problem: After a restore to both machines, we reboot both machines. This ensures all old connections to the database are removed. There are six process that can launch local and remote queries. This means that each machine will eventually be running 6 local and 6 remote sqlturbos. However, it frequently occurs that one of 6 remote sqlturbos freezes. In fact, I have seen cases where 4 of the 6 remote sqlturbos are frozen. Truss output indicates that the sqlturbo is waiting for a semaphor: # ps -ef | grep /usr/as/informix/lib/sqlturb root 626 412 0 Jul 18 ? 0:02 /usr/as/informix/lib/sqlturbo root 5.02.UC3 AAA#L780112 -nvaddb -l -g AAAAAk9OT root 922 412 0 Jul 19 ? 0:34 /usr/as/informix/lib/sqlturbo root 5.02.UC3 AAA#L780112 -nvaddb -l -g AAAAAk9OT root 920 412 0 Jul 19 ? 0:33 /usr/as/informix/lib/sqlturbo root 5.02.UC3 AAA#L780112 -nvaddb -l -g AAAAAk9OT root 921 412 0 Jul 19 ? 0:01 /usr/as/informix/lib/sqlturbo root 5.02.UC3 AAA#L780112 -nvaddb -l -g AAAAAk9OT root 919 412 0 Jul 19 ? 0:33 /usr/as/informix/lib/sqlturbo root 5.02.UC3 AAA#L780112 -nvaddb -l -g AAAAAk9OT root 926 412 0 Jul 19 ? 0:31 /usr/as/informix/lib/sqlturbo root 5.02.UC3 AAA#L780112 -nvaddb -l -g AAAAAk9OT root 13011 12993 5 14:35:13 pts0 0:00 grep /usr/as/informix/lib/sqlturb # truss -p 626 *** SUID: ruid/euid/suid = 0 / 0 / 96 *** *** SGID: rgid/egid/sgid = 1 / 96 / 96 *** read(0, 0x0015AA90, 1024) (sleeping...) # truss -p 920 *** SUID: ruid/euid/suid = 0 / 0 / 96 *** *** SGID: rgid/egid/sgid = 1 / 96 / 96 *** read(0, 0x0015AA90, 1024) (sleeping...) # truss -p 921 *** SUID: ruid/euid/suid = 0 / 96 / 96 *** *** SGID: rgid/egid/sgid = 1 / 96 / 96 *** semop(0, 0xEFFFEA88, 1) (sleeping...) # truss -p 919 *** SUID: ruid/euid/suid = 0 / 0 / 96 *** *** SGID: rgid/egid/sgid = 1 / 96 / 96 *** read(0, 0x0015AA90, 1024) (sleeping...) # truss -p 926 *** SUID: ruid/euid/suid = 0 / 0 / 96 *** *** SGID: rgid/egid/sgid = 1 / 96 / 96 *** read(0, 0x0015AA90, 1024) (sleeping...) Things we know about the problem: o Once all 6 remote sqlturbos are started there is no more freezing o Once all 6 remote sqlturbos have been running we don't see the problem again even if we shutdown until we have to restore the database o Powering the machine all the way down after a two shelf restore does not solve the problem Things we suspect about the problem: o Seems to occur most frequently when the data schema exists but contains no data. This may explains why we see the problem almost always after installation. Any new ideas would be appreciated! Cheers, Tim.