Re: More "IpcMemoryCreate: shmget failed" errors
От | The Hermit Hacker |
---|---|
Тема | Re: More "IpcMemoryCreate: shmget failed" errors |
Дата | |
Msg-id | Pine.BSF.4.21.0006161408430.722-100000@thelab.hub.org обсуждение исходный текст |
Ответ на | More "IpcMemoryCreate: shmget failed" errors (Sean Kelly <S.Kelly@newcastle.ac.uk>) |
Ответы |
Re: More "IpcMemoryCreate: shmget failed" errors
|
Список | pgsql-bugs |
Okay, my first question is is *what* OS are you running on? I get this sort of error under Solaris if I use a defautl install of both the OS and PostgreSQL, and have to use the -B option to reduce the number of segments that PostgreSQL wants to configure ... v6.5.3, I believe it was, changed the IPC usage so that it allocates all it needs when you start, vs growing as required, so that it pre-allocates ... so if you were using a pre-"this code" version and moving up, this might be triggering your problem ... On Fri, 16 Jun 2000, Sean Kelly wrote: > Hello again, > > A couple of weeks ago I reported a problem installing both 7.0 and > 7.0.2 where attempting to start "postmaster" produced an "IpcMemoryCreate: > shmget failed"-type error. > > Tom Lane suggested I use the "ipcclean" application to remove > the "old shmem segments". > > Running "ipcclean" with 7.0.2 gave "ipcclean: nothing removed" as > the result "but postmaster" still gave the same "IpcMemoryCreate: shmget > failed" error. > > So ... I deleted this 7.0.2 installation and decided to go back to > 6.5.3. This installed, but when trying to start "postmaster" I got: > > ----- - - ----- > IpcMemoryCreate: shmget failed (Permission denied) key=5432010, size=120, permission=700 > IpcMemoryAttach: shmat failed (Permission denied) id=130 > FATAL 1: AttachSLockMemory: could not attach segment > ----- - - ----- > > So, I ran ipcclean that comes with 6.5.3 which finished with no > messages, but "postmaster" still gives the "IpcMemoryCreate: shmget > failed". > > Hence, I cannot run PostgreSQL on this system any more :( > > Any suggestions? Is there a way that v7 could check for these old > segments from v6 and remove them? Is this even the problem here? Any > other ways to clear these suspected old segments at all? > > Thanks for your time, > > -- > Sean Kelly <S.Kelly@ncl.ac.uk> > "The best way to predict the future is to invent it" - Alan Kay > Marc G. Fournier ICQ#7615664 IRC Nick: Scrappy Systems Administrator @ hub.org primary: scrappy@hub.org secondary: scrappy@{freebsd|postgresql}.org
В списке pgsql-bugs по дате отправления: