Thread: Status of ipcclean

ipcclean(1) currently says:

| ipcclean cleans up shared memory and semaphore space from aborted backends
| by deleting all instances owned by user postgres. Only the DBA should
| execute this program as it can cause bizarre behavior (i.e., crashes) if
| run during multi-user execution. This program should be executed if
| messages such as semget: No space left on device are encountered when
| starting up the postmaster or the backend server.

AFAIR, with the 7.1 release the postmaster automatically recovers from
this situation.  Can someone come up with a better description of what
ipcclean is useful for, if there still is such a thing?

-- 
Peter Eisentraut   peter_e@gmx.net   http://funkturm.homeip.net/~peter

Peter Eisentraut <peter_e@gmx.net> writes:
> AFAIR, with the 7.1 release the postmaster automatically recovers from
> this situation.  Can someone come up with a better description of what
> ipcclean is useful for, if there still is such a thing?

I believe that ipcclean is no longer needed for preparing to start a new
postmaster.  It might possibly be useful if you wanted to clean up after
a dead postmaster that you did *not* intend to restart.

However, given the lack of portability and lack of robustness of the
script (including inability to deal with multiple-postmaster
situations), I think I'd vote for removing it altogether.
        regards, tom lane

Tom Lane writes:

> I believe that ipcclean is no longer needed for preparing to start a new
> postmaster.  It might possibly be useful if you wanted to clean up after
> a dead postmaster that you did *not* intend to restart.
>
> However, given the lack of portability and lack of robustness of the
> script (including inability to deal with multiple-postmaster
> situations), I think I'd vote for removing it altogether.

Can other people voice their opinions what to do with ipcclean?

-- 
Peter Eisentraut   peter_e@gmx.net   http://funkturm.homeip.net/~peter