Child process during shellcommand hangs regardless of timeout...

[Christian Pearce]

Here is a quick example script that causes the problem...

control:
    actionsequence = ( shellcommands )

    domain = ( commnav.com )

shellcommands:

    "/bin/rpm -Uvvh /usr/local/tar/webmin-0.87.rpm"

    timeout=90

webmin start a httpd style service during the postinstall which
listens TCP and UDP port 10000.  Here is output from a strace

[pid  5361] time(NULL)                  = 996774943
[pid  5361] wait4(-1, 0xbffff538, WNOHANG, NULL) = -1 ECHILD (No child
processes)
[pid  5361] lseek(6, 0, SEEK_SET)       = 0
[pid  5361] read(6, "\0\0\366\3\357\3\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,
1024) = 1024
[pid  5361] read(6, "", 1024)           = 0
[pid  5361] select(8, [4 5], NULL, NULL, {10, 0}) = 0 (Timeout)
[pid  5361] time(NULL)                  = 996774953
[pid  5361] wait4(-1, 0xbffff538, WNOHANG, NULL) = -1 ECHILD (No child
processes)
[pid  5361] lseek(6, 0, SEEK_SET)       = 0
[pid  5361] read(6, "\0\0\366\3\357\3\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"...,
1024) = 1024
[pid  5361] read(6, "", 1024)           = 0
[pid  5361] select(8, [4 5], NULL, NULL, {10, 0}) = 0 (Timeout)

and here is the output from lsof...
miniserv. 5361    root    4u  IPv4     259817                TCP
*:ndmp (LISTEN)
miniserv. 5361    root    5u  IPv4     259818                UDP
*:ndmp

does anyone have any clues as to why the timeout doesn't work?