Unable to kill some sessions


I am unable to kill two sessions that iimonitor marks them as:

cs_state: CS_COMPUTABLE cs_mask: CS_DEAD_MASK,CS_IRPENDING_MASK

I have tried:

- Kill them with ipm. After it, ipm informed that session was removed,
but it was not. Still appeared in the session list.

- Kill the process (in the remote machine) that created them. In this
case a JDBC server. No change.

- Kill them with iimonitor. No result. I also tried the
suspend,remove,resume trick with no results.

I suspect that they perhaps are comsuming quite CPU in a indefinite
loop or something as the system is quiet but CPU is unusually high.

(iimonitor shows that the query is "open ~Q cursor for INGSTMT0 for
readonly using ~V , ~V , ~V , ~V , ~V , ~V")

Any ideas?

Thanks in advance


Carlos

.

Relevant Pages

  • Re: unkillable sids
    ... same sql  and the onstat -g ses are all the same. ... but the result is usually just the stack of what the thread ... the onmode command to kill sessions doesn't actually ... kill the session, it just sets a kill flag that needs to be checked ...
    (comp.databases.informix)
  • Re: Unable to kill some sessions
    ... > I am unable to kill two sessions that iimonitor marks them as: ... > - Kill them with ipm. ... Still appeared in the session list. ... > loop or something as the system is quiet but CPU is unusually high. ...
    (comp.databases.ingres)
  • RE: [Info-ingres] Unable to kill some sessions
    ... I am unable to kill two sessions that iimonitor marks them as: ... Still appeared in the session list. ... I suspect that they perhaps are comsuming quite CPU in a indefinite ...
    (comp.databases.ingres)
  • Re: problem connecting to db.
    ... rerun the service with above problem, and kill its oracle-session ... alter system kill session ... ... So do you see the memory being released after you kill? ... like lock type we might be able to help figure out what is going on. ...
    (comp.databases.oracle.misc)
  • Re: unkillable sids
    ... same sql and the onstat -g ses are all the same. ... but the result is usually just the stack of what the thread ... the onmode command to kill sessions doesn't actually ... kill the session, it just sets a kill flag that needs to be checked ...
    (comp.databases.informix)