Tom Lane wrote:
> Alvaro Herrera <alvherre@commandprompt.com> writes:
> > How about this?
>
> I think the accounting for the AV launcher in shmem allocation is a bit
> confused yet --- for instance, isn't MaxBackends already including the
> launcher? I wonder if it would be cleaner to include the launcher in
> the autovacuum_max_workers parameter, and increase the min/default
> values of that by one.
Huh, yeah, sorry about that -- fixed here. I think the name of the
param, which includes "worker", precludes from raising the values.
Changes between v2 and v3:
diff -u src/backend/storage/lmgr/proc.c src/backend/storage/lmgr/proc.c
--- src/backend/storage/lmgr/proc.c 31 Aug 2009 13:36:56 -0000
+++ src/backend/storage/lmgr/proc.c 31 Aug 2009 16:14:08 -0000
@@ -103,7 +103,7 @@
/* AuxiliaryProcs */
size = add_size(size, mul_size(NUM_AUXILIARY_PROCS, sizeof(PGPROC)));
/* MyProcs, including autovacuum workers and launcher */
- size = add_size(size, mul_size(MaxBackends + 1, sizeof(PGPROC)));
+ size = add_size(size, mul_size(MaxBackends, sizeof(PGPROC)));
/* ProcStructLock */
size = add_size(size, sizeof(slock_t));
@@ -192,6 +192,7 @@
ProcGlobal->freeProcs = &procs[i];
}
+ /* note: the "+1" here accounts for the autovac launcher */
procs = (PGPROC *) ShmemAlloc((autovacuum_max_workers + 1) * sizeof(PGPROC));
if (!procs)
ereport(FATAL,
diff -u src/backend/utils/misc/guc.c src/backend/utils/misc/guc.c
--- src/backend/utils/misc/guc.c 31 Aug 2009 03:07:47 -0000
+++ src/backend/utils/misc/guc.c 31 Aug 2009 16:12:56 -0000
@@ -7570,7 +7570,7 @@
static bool
assign_maxconnections(int newval, bool doit, GucSource source)
{
- if (newval + autovacuum_max_workers > INT_MAX / 4)
+ if (newval + autovacuum_max_workers + 1 > INT_MAX / 4)
return false;
if (doit)
--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
