Tom Lane wrote:
> Alvaro Herrera <alvherre(at)commandprompt(dot)com> writes:
>> How about getting ShareUpdateExclusiveLock on manual analyze and plain
>> AccessShareLock on autovacuum-induced analyze?
> Wouldn't fix the original problem because those two lock types don't
> conflict; hence might as well keep the behavior simple.
What about a new separate lock type for analyze? Couldn't that really
solve the issue? I know I'm just hand-waving here ;-)
In response to
pgsql-hackers by date
|Next:||From: Michael Paesold||Date: 2007-10-02 09:23:12|
|Subject: Re: First steps with 8.3 and autovacuum launcher|
|Previous:||From: Magnus Hagander||Date: 2007-10-02 09:18:23|
|Subject: Re: [COMMITTERS] pgsql: Use BIO functions to avoid passing FILE * pointers to OpenSSL|