Re: pgsql: Set MaxBackends only on bootstrap and standalone modes

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>
Cc: pgsql-committers(at)postgresql(dot)org
Subject: Re: pgsql: Set MaxBackends only on bootstrap and standalone modes
Date: 2013-01-02 21:31:25
Message-ID: 26055.1357162285@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-committers

Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> writes:
> Set MaxBackends only on bootstrap and standalone modes
> ... not on auxiliary processes.

Uh, why is that a good idea? bootstrap mode, for instance, thinks it's
an auxiliary process, and is still broken.

(gdb) b errfinish
Breakpoint 1 at 0x73c8d0: file elog.c, line 393.
(gdb) run --boot
Starting program: /home/tgl/testversion/bin/postgres --boot

Breakpoint 1, errfinish (dummy=0) at elog.c:393
393 {
Missing separate debuginfos, use: debuginfo-install glibc-2.16-28.fc18.x86_64
(gdb) bt
#0 errfinish (dummy=0) at elog.c:393
#1 0x0000000000747c6e in hash_search_with_hash_value (hashp=0xbfa1f0,
keyPtr=keyPtr(at)entry=0x8ce7d0 <ScratchTargetTag>,
hashvalue=<optimized out>, action=action(at)entry=HASH_ENTER,
foundPtr=foundPtr(at)entry=0x0) at dynahash.c:958
#2 0x0000000000747eb9 in hash_search (hashp=<optimized out>,
keyPtr=keyPtr(at)entry=0x8ce7d0 <ScratchTargetTag>,
action=action(at)entry=HASH_ENTER, foundPtr=foundPtr(at)entry=0x0)
at dynahash.c:805
#3 0x000000000066fd4d in InitPredicateLocks () at predicate.c:1130
#4 0x000000000065cb25 in CreateSharedMemoryAndSemaphores (
makePrivate=makePrivate(at)entry=1 '\001', port=port(at)entry=0) at ipci.c:207
#5 0x000000000074a1f9 in InitCommunication () at postinit.c:388
#6 BaseInit () at postinit.c:460
#7 0x00000000004d7f8d in AuxiliaryProcessMain (argc=1, argc(at)entry=2,
argv=0xbf8dd8, argv(at)entry=0xbf8dd0) at bootstrap.c:362
#8 0x0000000000458bdf in main (argc=2, argv=0xbf8dd0) at main.c:191
(gdb) p MaxBackends
$1 = 0
(gdb)

> Hopefully this will turn the buildfarm green again, and put an end to
> today's silliness.

Not there yet, either in the buildfarm or locally for me. Maybe
you should revert this whole change and come at it fresh tomorrow.

regards, tom lane

In response to

Browse pgsql-committers by date

  From Date Subject
Next Message Alvaro Herrera 2013-01-02 21:40:51 pgsql: Fix IsUnderPostmaster/EXEC_BACKEND confusion
Previous Message Alvaro Herrera 2013-01-02 21:00:17 pgsql: Set MaxBackends only on bootstrap and standalone modes