DD process restart failures
Hi all,
ts5.5.2 + patches + sp4, dd 5.6.0 Build 19499 Interwoven 20030506, solairs 7
I'm seeing intermittent DAS restarts overnight for log rotation. By stopping dd with iwsyncdb.ipl or /etc/init.d/iwdd stop, I see the process bound to port 2345 come up fine, but the perl process does not come back up. A truss of the running perl process while it's quiet, shows the following on a predictable basis:
<snip>
poll(0xFEC07A88, 0, 10) = 0
poll(0xFEC07A88, 0, 10) = 0
Received signal #14, SIGALRM, in lwp_sema_wait() [caught]
lwp_sema_wait(0xFF38E4C8) Err#91 ERESTART
sigprocmask(SIG_SETMASK, 0xFED09E28, 0x00000000) = 0
lwp_sema_post(0xFEE0BE78) = 0
sigprocmask(SIG_SETMASK, 0xFF399810, 0x00000000) = 0
lwp_sema_wait(0xFEE0BE78) = 0
setcontext(0xFED09718)
sigprocmask(SIG_BLOCK, 0xFF38E4B8, 0x00000000) = 0
lwp_sema_post(0xFF38E4C8) = 0
setitimer(ITIMER_REAL, 0xFED09CB8, 0x00000000) = 0
sigprocmask(SIG_UNBLOCK, 0xFF38E4B8, 0x00000000) = 0
lwp_sema_wait(0xFF38E4C8) = 0
sigprocmask(SIG_BLOCK, 0xFF38E4B8, 0x00000000) = 0
setitimer(ITIMER_REAL, 0xFED09CB8, 0x00000000) = 0
sigprocmask(SIG_UNBLOCK, 0xFF38E4B8, 0x00000000) = 0
poll(0xFEC07A88, 0, 10) = 0
poll(0xFEC07A88, 0, 10) = 0
<snip>
I've attached the truss of iwsyncdb.ipl -stopddd and init.d/iiwdd stop as it's too big for this format. There are some errors, but I think they are normal as they pertain to stats of libraries which cannot be found in $IWHOME/tools/java1.3/jre/ or some subdirs, but are found elsewhere in $DDHOME/lib. The library is libscijni.so.
As I indicated the failures are intermittent. The $DDHOME/install/dd_pid file is writeable, but is only updated during a successful restart.
Thanks in advance for any comments,
R.Barger
CSFS TOP
Zurich
