Bruce Momjian <bruce@momjian.us> writes:
> Tom Lane wrote:
>> There actually were two smgr storage modules in the code we inherited
>> from Berkeley, and I think there were probably more at one time.
> Yes, the second storage manager we had was for WORM drives, or more
> accurately, stubs were left in our code for WORM drives.
No, the other storage manager in the Berkeley sources was for keeping
stuff in "hopefully stable" main memory:
http://anoncvs.postgresql.org/cvsweb.cgi/pgsql/src/backend/storage/smgr/Attic/mm.c?rev=1.1;content-type=text%2Fplain
Of course that was dead as a doornail after people stopped using
magnetic-core main memory. There are references to a WORM storage
manager in the Berkeley Postgres papers, but it's not clear to me
if they ever wrote one or merely theorized about it. The hint-bit
stuff, to take just one example, would never have worked on WORM
storage, and yet it's implemented above the smgr API. So it seems
kinda unlikely that there was ever a working WORM smgr.
regards, tom lane
