polld still dying

[angelb]

> Alright, I'm kind of running out of ideas why cups-polld is killed
> while polling. The following excerpts are from truss and error_log
> showing the cups-polld daemon running and then receiving a SITERM
> 15(PID 770072):
>
>  +8296  757982: kread(3, " D E B U G :  \0\0\0\0\0".., 1023)    = 7
>  +8297  757982: kwrite(1, " E   [ 3 0 / J u n / 2 0".., 86)     = 86
>  +8298  757982: 6.8534:        close(3)                         = 0
>  +8299  757982: 6.8540:        close(3)                         Err#9  EBADF
>  +8300  757982: kwrite(1, " d   [ 3 0 / J u n / 2 0".., 78)     = 78
>  +8301  757982: 6.8551:        kill(770072, 15)                 = 0
>  +8302  770072: kwrite(2, " D E B U G :  ", 7)                  = 7
>  +8303  770072:     Received signal #15, SIGTERM [default]
>  +8304  757982: 6.8556:        _select(4096, 0x20008328, 0x2000A338, 0x00000000, 0x2FF1E828) Err#4  EINTR
>  +8305  757982:     Received signal #20, SIGCHLD [caught]
>  +8306  757982: 6.8581:        ksetcontext_sigreturn(0x2FF1E3D0, 0x00000000, 0x2FF1E3D0, 0x2FF3B000, 0x100018F8, 0x0000D0B2, 0x20ADDA00, 0x2FF39F28)
>  +8307  757982: kwrite(1, " d   [ 3 0 / J u n / 2 0".., 50)     = 50
>  +8308  757982: 6.8591:        kwaitpid(0x2FF1E390, -1, 5, 0x00000000, 0x00000000) = 770072
>  +8309  757982: kwrite(1, " E   [ 3 0 / J u n / 2 0".., 108)    = 108
>  +8310  757982: 6.8604:        kwaitpid(0x2FF1E390, -1, 5, 0x00000000, 0x00000000) Err#10 ECHILD
>  +8311  770072: *** process killed ***
>  +8312  757982: 6.8608:        _select(4096, 0x20008328, 0x2000A338, 0x00000000, 0x2FF1E828) = 0
>
>
> d [30/Jun/2006:14:50:14 +0600] cupsdStartProcess("/opt/TWWfsw/cups121/lib/daemon/cups-polld", 2ff1e1c0, 2ff1e1d8, -1, -1, 4)
> D [30/Jun/2006:14:50:15 +0600] cupsdStartPolling: Started polling daemon for occam503:631, pid = 770072
> ...
> d [30/Jun/2006:14:50:16 +0600] cupsdDeletePrinter(p=20042168(AR0J), update=1)
> d [30/Jun/2006:14:50:16 +0600] stringpool: 168 strings, 5192 allocated, 3992 total bytes
> d [30/Jun/2006:14:50:16 +0600] select_timeout: 86400 seconds to do nothing
> E [30/Jun/2006:14:50:16 +0600] cupsdUpdatePolling: all polling processes have exited!
> d [30/Jun/2006:14:50:16 +0600] cupsdStopPolling: removing fd 3 from InputSet.
> d [30/Jun/2006:14:50:16 +0600] process_children()
> E [30/Jun/2006:14:50:16 +0600] PID 770072 (/opt/TWWfsw/cups121/lib/daemon/cups-polld) crashed on signal 15!
> d [30/Jun/2006:14:50:17 +0600] select_timeout: 86400 seconds to do nothing
>
> Anyone else running CUPS 1.2.1 on RedHat and AIX or Solaris and not
> having this issue?
>
> The other client running CUPS 1.1.23 works okay with the CUPS server
> running 1.2.1 but the client running 1.2.1 is just not working. It
> starts to poll up to 2 printers and then killed. How's that?
>
> Any ideas or suggestion is appreciated.

Just a follow-up.

When running cups-polld by itself, it seems to poll all the printers
okay. However, when it's forked off or called by cupsd, then it's
killed by the parent. I'm beginning to suspect the scheduler on 1.2.1
itself is causing this because 1.1.23 polls just fine.

I've also compared the environment and variables between the running
cupsd and cups-polld in truss and don't see anything suspicious.

Thanks,
Angel