Hi all
We are building hosting with apache + php ( our own mod_virtual module )
with about 10.000 wirtul domains + PostgreSQL.
PostgreSQL is on a different machine ( 2 x intel xeon 2.4GHz 1GB RAM
scsi raid 1+0 )
I've made some test's - 3000 databases and 400 clients connected at same
time. These
W liście z pią, 31-10-2003, godz. 11:52, Matt Clark pisze:
> > I could made persistent connection, but with 10.000 clients it will kill
> > the server.
>
> But if they're virtual domains, why would you need one connection per domain? You
> should only need one connection per apache
> process...
W liście z pią, 31-10-2003, godz. 12:25, Matt Clark pisze:
> Ooh, I see. That's a tricky one. Do you really need that level of separation?
Well, if you talk with the clients, and they promise, that they will not
access to other databasess, and specially don't do "drop database
my_bes_fried_db"
W liście z pią, 31-10-2003, godz. 13:33, Matt Clark pisze:
> > W liście z pią, 31-10-2003, godz. 12:25, Matt Clark pisze:
> > > Ooh, I see. That's a tricky one. Do you really need that level of separation?
> >
> > Well, if you talk with the clients, and they promise, that they will not
> > acce
W liście z pią, 31-10-2003, godz. 13:06, Gaetano Mendola pisze:
> Marek Florianczyk wrote:
>
> > But my problem is that when I hit command:
> > psql -h 127.0.0.1 dbname dbuser
> > I'm waiting about 3-5 sec to enter psql monitor, so every new connection
> > fro
W liście z pią, 31-10-2003, godz. 13:54, Jamie Lawrence pisze:
> On Fri, 31 Oct 2003, Matt Clark wrote:
>
> > I was more thinking that it might be possible to manage the security at a
> > different level than the DB.
> >
>
>
> We do this with users and permissions.
>
> Each virtual host has
W liście z pią, 31-10-2003, godz. 15:23, Tom Lane pisze:
> Marek Florianczyk <[EMAIL PROTECTED]> writes:
> > We are building hosting with apache + php ( our own mod_virtual module )
> > with about 10.000 wirtul domains + PostgreSQL.
> > PostgreSQL is on a different mach
-)
>
> M
>
> > -----Original Message-
> > From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of Marek Florianczyk
> > Sent: 31 October 2003 13:20
> > To: Jamie Lawrence
> > Cc: Matt Clark; [EMAIL PROTECTED]
> > Subject: Re: [ADMIN] performanc
W liście z pią, 31-10-2003, godz. 16:51, Mike Rylander pisze:
> On Friday 31 October 2003 09:59 am, Marek Florianczyk wrote:
> > W liście z pią, 31-10-2003, godz. 15:23, Tom Lane pisze:
> > > Marek Florianczyk <[EMAIL PROTECTED]> writes:
> > > > We are building
Hi all
I was asking for help, a week ago. Performace tests took mi some more
time because of other things that I had to do.
The problem was: tune PostgreSQL to work with 10.000 databases.
Tom Lane (thanks) suggested solution: one database and 10.000 schemas.
>From now I will write switch "database
W liście z śro, 05-11-2003, godz. 14:48, Jeff pisze:
> On 05 Nov 2003 14:33:33 +0100
> Marek Florianczyk <[EMAIL PROTECTED]> wrote:
>
> >
> > During this test I was changing some parameters in postgres, and send
> > kill -HUP ( pg_ctl reload ). I still don'
W liście z śro, 05-11-2003, godz. 17:24, Andrew Sullivan pisze:
> On Wed, Nov 05, 2003 at 08:48:52AM -0500, Jeff wrote:
> >
> > as far as I know, -HUP won't make things like shared buffer changes
> > take. you need a full restart of PG.
>
> It definitely will not. Anything that can only be set
W liście z śro, 05-11-2003, godz. 17:18, Bruno Wolff III pisze:
> On Wed, Nov 05, 2003 at 16:14:59 +0100,
> Marek Florianczyk <[EMAIL PROTECTED]> wrote:
> > One database with 3.000 schemas works better than 3.000 databases, but
> > there is REAL, BIG problem, and I w
W liście z śro, 05-11-2003, godz. 18:59, Tom Lane pisze:
> Marek Florianczyk <[EMAIL PROTECTED]> writes:
> > Each client was doing:
>
> > 10 x connect,"select * from table[rand(1-4)] where
> > number=[rand(1-1000)]",disconnect--(fetch one row)
>
>
W liście z śro, 05-11-2003, godz. 19:23, Jeff pisze:
> On 05 Nov 2003 19:01:38 +0100
> Marek Florianczyk <[EMAIL PROTECTED]> wrote:
>
> > and it works better, but no revelation, when I do "\d
> > schemaname.table" it's better. I've to still wait abo
W liście z śro, 05-11-2003, godz. 19:34, Tom Lane pisze:
> Marek Florianczyk <[EMAIL PROTECTED]> writes:
> > But did you do that under some database load ? eg. 100 clients
> > connected, like in my example ? When I do these queries "\d" without any
> > cli
W liście z śro, 05-11-2003, godz. 19:52, Tom Lane pisze:
> Marek Florianczyk <[EMAIL PROTECTED]> writes:
> > Maybe reconnect is to often, but how to explain that reular queries like
> > select * from table1 ale much faster than \d's ? ( my post to Jeff )
>
> [
> Heck, you're already pushing the performance envelope with 3,000 users,
> might as well go for the faster of the two and you'll have one less
> scheduled upgrade ahead of you.
>
> When do you need to go live? If it's >1 month, then I'd definitely
> recommend 7.4.
heh... ;)
PostgreSQL 7.4
W liście z czw, 06-11-2003, godz. 15:37, Jeff pisze:
> On 06 Nov 2003 15:21:03 +0100
> Marek Florianczyk <[EMAIL PROTECTED]> wrote:
>
>
> > fsync = false
>
> HOLD THE BOAT THERE BATMAN!
>
> I would *STRONGLY* advise not running wi
W liście z pią, 28-11-2003, godz. 13:51, Oli Sennhauser pisze:
> Hello
>
> We are planning and offering a low-cost Photo-Archive in TB range. Is
> there any experience available with that amount of Data (1-3 TB)?
> What kind of machines (CPU, RAM)? What kind of I/O-System? What kind of
> Networ
W liście z pon, 15-12-2003, godz. 16:30, Rich Cullingford pisze:
> All,
> I did a pg_dumpall in preparation for moving one of our databases from
> PG7.3 to PG7.4, but I just realized I have another problem: that DB
> (which has served us faithfully for some time) was created for superuser
> 'po
21 matches
Mail list logo
