Re: pg_prewarm

<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px;
-qt-user-state:0;">Lemercredi 20 juin 2012 21:53:43, Peter Eisentraut a écrit :<p style=" margin-top:0px;
margin-bottom:0px;margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">> On
tis,2012-04-10 at 13:29 +0200, Cédric Villemain wrote:<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px;
margin-right:0px;-qt-block-indent:0; text-indent:0px; -qt-user-state:0;">> > I have no problem deprecating
overlappingfeatures from pgfincore as<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px;
-qt-block-indent:0;text-indent:0px; -qt-user-state:0;">> > soon as I can do a «depend:pg_prewarm[os_warm]» :)
...Itwould have<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0;
text-indent:0px;-qt-user-state:0;">> > been better to split pgfincore analyze and warming parts times<p style="
margin-top:0px;margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px;
-qt-user-state:0;">>> ago, anyway.<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px;
margin-right:0px;-qt-block-indent:0; text-indent:0px; -qt-user-state:0;">> <p style=" margin-top:0px;
margin-bottom:0px;margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">> So
pg_prewarmis now pending in the commitfest, so let's see what the<p style=" margin-top:0px; margin-bottom:0px;
margin-left:0px;margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">> situation is.<p
style="-qt-paragraph-type:empty;margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px;
-qt-block-indent:0;text-indent:0px; "> <p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px;
-qt-block-indent:0;text-indent:0px; -qt-user-state:0;">I have refused to propose pgfincore so far because BSD didn't
supportedPOSIX_FADVISE (but already supported mincore(2)).<p style=" margin-top:0px; margin-bottom:0px;
margin-left:0px;margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">Now, things change and
pgfincoreshould work on linux, bsd, hp, ... (but not on windows)<p style=" margin-top:0px; margin-bottom:0px;
margin-left:0px;margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">I'll be happy to propose it
ifpeople want.<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px;
-qt-block-indent:0;text-indent:0px; "> <p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px;
-qt-block-indent:0;text-indent:0px; -qt-user-state:0;">> I'm worried about the overlap with pgfincore. It's pretty
well<pstyle=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0;
text-indent:0px;-qt-user-state:0;">> established in this space, and has about 73% feature overlap with<p style="
margin-top:0px;margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px;
-qt-user-state:0;">>pg_prewarm, while having more features all together. So it would seem<p style=" margin-top:0px;
margin-bottom:0px;margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">> to me
thatit would be better to add whatever is missing to pgfincore<p style=" margin-top:0px; margin-bottom:0px;
margin-left:0px;margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">> instead. Or split
pgfincore,as suggested above, but that would upset<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px;
margin-right:0px;-qt-block-indent:0; text-indent:0px; -qt-user-state:0;">> existing users. But adding severely
overlappingstuff for no technical<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px;
-qt-block-indent:0;text-indent:0px; -qt-user-state:0;">> reasons (or there any?) doesn't sound like a good idea.<p
style="-qt-paragraph-type:empty;margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px;
-qt-block-indent:0;text-indent:0px; "> <p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px;
-qt-block-indent:0;text-indent:0px; -qt-user-state:0;">And I am also worried with the overlap.<p
style="-qt-paragraph-type:empty;margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px;
-qt-block-indent:0;text-indent:0px; "> <p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px;
-qt-block-indent:0;text-indent:0px; -qt-user-state:0;">> Also, Robert has accurately described this as "mechanism,
notpolicy".<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0;
text-indent:0px;-qt-user-state:0;">> That's fine, that's what all of our stuff is. Replication and most of<p style="
margin-top:0px;margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px;
-qt-user-state:0;">>postgresql.conf suffers from that. Eventually someone will want to<p style=" margin-top:0px;
margin-bottom:0px;margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">>
createa way to save and restore various caches across server restarts,<p style=" margin-top:0px; margin-bottom:0px;
margin-left:0px;margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">> as discussed before.
Wouldthat mean there will be a third way to do<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px;
margin-right:0px;-qt-block-indent:0; text-indent:0px; -qt-user-state:0;">> all this, or could we at least align
thingsa bit so that such a<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px;
-qt-block-indent:0;text-indent:0px; -qt-user-state:0;">> facility could use most of the proposed prewarming stuff?
(Patchesfor<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0;
text-indent:0px;-qt-user-state:0;">> the cache restoring exist, so it should be possible to predict this a<p style="
margin-top:0px;margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px;
-qt-user-state:0;">>little bit.)<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px;
margin-left:0px;margin-right:0px; -qt-block-indent:0; text-indent:0px; "> <p style=" margin-top:0px; margin-bottom:0px;
margin-left:0px;margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">It makes some time I have a
lookat the postgresql source code about readBuffer and friends. AFAIR pgfincore needed some access to file decsriptor
whichwere not possible with PostgreSQL core functions.<p style="-qt-paragraph-type:empty; margin-top:0px;
margin-bottom:0px;margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; "> <p style=" margin-top:0px;
margin-bottom:0px;margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">I can have
alook as this is near the stuff I'll work on next (posix_fadvice random/sequential/normal applyed directly by
PostgreSQL,instead of relying on hacks around read-the-first-block to start readahead).<p
style="-qt-paragraph-type:empty;margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px;
-qt-block-indent:0;text-indent:0px; "> <p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px;
-qt-block-indent:0;text-indent:0px; -qt-user-state:0;">-- <p style=" margin-top:0px; margin-bottom:0px;
margin-left:0px;margin-right:0px; -qt-block-indent:0; text-indent:0px; -qt-user-state:0;">Cédric Villemain +33 (0)6 20
3022 52<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0;
text-indent:0px;-qt-user-state:0;">http://2ndQuadrant.fr/<p style=" margin-top:0px; margin-bottom:0px; margin-left:0px;
margin-right:0px;-qt-block-indent:0; text-indent:0px; -qt-user-state:0;">PostgreSQL: Support 24x7 - Développement,
Expertiseet Formation<p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px;
margin-right:0px;-qt-block-indent:0; text-indent:0px; ">