Re: linux-kernel say bug

[Greg Harper]

I get the same error, but if you look at the next line: Dec  4 13:15:25 mail kernel: application bug: dumper(12282) has SIGCHLD set to SIG_IGN but calls wait(). Dec  4 13:15:25 mail kernel: (see the NOTES section of 'man 2 wait'). Workaround activated. Then look at the  man page on wait: NOTES       The Single Unix Specification describes a flag SA_NOCLDWAIT  (not  sup-       ported under Linux) such that if either this flag is set, or the action       for SIGCHLD is set to SIG_IGN then children that  exit  do  not  become       zombies and a call to wait() or waitpid() will block until all children       have exited, and then fail with errno set to ECHILD.       The original POSIX standard left the behaviour of  setting  SIGCHLD  to       SIG_IGN  unspecified.   Later  standards,  including  SUSv2  and  POSIX       1003.1-2001 specify the behaviour just described as  an  XSI-compliance       option.   Linux  does  not conform to the second of the two points just       described: if a wait() or waitpid() call is made while SIGCHLD is being       ignored, the call behaves just as though SIGCHLD were not being ignored,       that is, the call blocks until  the  next  child  terminates  and  then       returns the PID and status of that child. So its really nothing serious, just something that should be fixed in future versions. -Greg Mats Blomstrand wrote: modified I don't recall. Is your's an older amanda release? On that machine its amanda-2.4.4. //Mats