Re: How to tell who sent KILL signal

From: Pascal Bourguignon (spam_at_thalassa.informatimago.com)
Date: 10/10/03

  • Next message: David Schwartz: "Re: How to build this time format?" 
    Date: 10 Oct 2003 07:07:07 +0200

    "Chris Markle" <cmarkle@sendmail.com> writes:

    > Hi - We're seeing a problem where when we start on application (happens to
    > be a sieve implementation) and it starts handling mail, another unrelated
    > (mostly) application (happens to be imap server, but sieve is talking to
    > lmtp server not imap) gets a SIGKILL delivered to it. Supposedly a truss
    > shows the KILL being delivered to the one process but not being sent from
    > the first. Any thoughts about how we can determine who or what is sending
    > the KILL signal into the unrelated process? Chris

    AFAIK, that is not possible.

    The kernel would need to keep a list of signaling pid
    (imagine: while true ; do kill -KILL $PID & done )

    In addition, some signals and in particular SIGKILL can be sent by the
    kernel itself. (What about a killing PID = 0 ?) 

    -- 
__Pascal_Bourguignon__
http://www.informatimago.com/
Do not adjust your mind, there is a fault in reality.
  • Next message: David Schwartz: "Re: How to build this time format?"

    Relevant Pages

    • Re: How to tell who sent KILL signal
      ... > lmtp server not imap) gets a SIGKILL delivered to it. ... Supposedly a truss ... > the KILL signal into the unrelated process? ... The kernel would need to keep a list of signaling pid ...
      (comp.unix.solaris)
    • Re: How to tell who sent KILL signal
      ... >> the KILL signal into the unrelated process? ... > The kernel would need to keep a list of signaling pid ...
      (comp.unix.programmer)
    • Re: How to tell who sent KILL signal
      ... >> the KILL signal into the unrelated process? ... > The kernel would need to keep a list of signaling pid ...
      (comp.unix.solaris)