| |
| |
| |
|
Page: 1 2 3 4 5
Comments:
<0> apparently PG 8.1 does SSL connections now?
<1> just 1
<1> a few other minor projects consume a few dbs as well
<1> but the major thing is the school management
<2> i need larger scale.. it needs to be proven, venture capitalists won't go for pg.. that's all i'm saying it can't be sold ebcause there is no proof it works for large scale web stuff with lots of reads
<2> http://www.postgresql.org/about/casestudies/ looks like a joke compared to http://www.mysql.com/why-mysql/case-studies/
<0> then use My
<1> Go for mysql then
<0> if you aren't going to take our advice, then don't
<1> I find postgre handles larger server loads better
<1> at least it dies gracefully... mysql just chokes and dies
<2> can it be optimized better than mysql for lots of reads?
<2> for fast reads
<0> inf00, PG stomps My
<1> inf00: its slightly slower afaik
<1> but it handles larger loads
<2> tech are there any papers out there that you are aware of on that? i'm just looking for reliable printed info to go on
<1> inf00: google is your best friend
<0> it's mine
<3> Use DB2.
<2> i did google
<1> Oracle it
<3> **** oracle
<3> DB2 pwnz oracle
<2> not going commercial...
<4> just how different is the syntax of pg from that of my?
<3> Use the free version of DB2.
<0> dude, it's SQL
<1> I was using oracle previously before pg
<0> almost no difference as far as queries goes
<1> Robdgreat: not that great...
<3> Robdgreat: PG is stricter. For example "Stuff" is a field name and 'Stuff' is a string.
<4> fair enough
<0> the system stuff is way different though
<0> you dont "use" databases
<1> Its just that in php ... mysql has neater stuff
<0> PG's field names are case sensitive
<1> mysql_insert_id is one thing that I really missed
<2> DjKaye such as?
<2> oh
<1> till i found out about oids
<4> I actually used that for the first time today
<1> :/
<0> yeah
<0> the keys are in a different table
<0> i'm still not sure why that is
<0> but i'm ***uming there's a good reason for it
<4> I'll learn more tomorrow. I'm having trouble keeping my eyes open
<1> inf00: nothing beats first hand experience
<1> Give both a shot for stress testing
<0> yeah info, fire one upt
<0> up
<4> g'night, fools
<0> later rob
<1> btw, one reason I moved to postgres from mysql
<1> was stored procedures and transactions
<3> yeah
<2> Dj.. i would but pg is a pain to setup and get to know and it takes time to test.. i wish there were good papers out on pg
<3> and triggers.
<3> Which 5.0 has, but they still pretty well ****.
<1> inf00: if wishes were pennies, I'd be friggin *** rich
<1> =D
<1> lets just say that mysql is prolly easier to use
<3> That's true
<1> but pg is more like a free version of oracle
<3> It IS easier to use.
<3> and that's true too
<3> in fact it's so close
<3> that someone took PG's code base
<3> ripped out the SQL parser
<3> and added one that's perfectly Oracle compliant.
<3> and they sell it now.
<3> for way cheaper than oracle
<2> what do they call it?
<3> and it works with anything oracle out of the box
<3> lemme find it
<3> EnterpriseDB
<3> (lame name)
<1> bleah
<2> heh
<1> Anyway inf00: just code your apps with a proper abstraction layer
<1> make you sure you don't have any db specific queries
<3> http://www.enterprisedb.com/products/index.do
<1> and you're just about set
<0> info, you could use pear to do your DB connect
<1> or adodb
<0> and then you could switch between My and PG seamlessly
<2> .do = oracle ?
<1> which I happen to use pretty frequently
<0> that's what i did
<1> Technobabble: almost...
<3> I don't know what .do is
<1> You're still missing the mysql_insert_id thingie
<1> heh
<2> yeah i use my own db abstraction
<0> yeah
<0> that's true
<3> not really
<3> just read the sequence ***ociated with it
<3> in the same transaction
<1> the query's different tho
<0> zen, why does PG not do the sql ID thing
<1> select * from table where oid ='oid';
<3> because it's smarter
<3> and uses sequences
<0> well i figured there was a good reason
<3> that can be arbitrary values
<0> but i don't understand it
<2> what's this oid thing?
<3> oid's are obsolete
<1> object identified
<1> object identifier
<1> pardon moi
<3> they were an old way to identify objects in the database
<3> but are gone now
<3> well mostly
<0> wow
<0> gaim has a spell checker for esperanto
<3> Technobabble: there isn't a pgsql_last_id because there is no such thing as an "id". There are sequences on columns, and you can put more than one sequence per row.
<3> You simply read the sequence for the column you want in the same transaction and voila, same results.
<3> You can even have a user-defined sequence
<3> that produces say, even numbers only
<3> I think 8.0+ even supports 'char' and 'varchar' sequences.
<1> Hmm ..
<1> I have yet to make the move to 8
<2> psql -- sql: could not connect to server: No such file or directory Is the server running locally and accepting connections on Unix domain socket "/tmp/.s.PGSQL.5432"?
<1> ps -x
<1> =.='
<3> psql -h localhost
<2> same thing
<2> TCP/IP connections on port 5432? double checking iptables
<1> bleh
<1> I'm out...
<2> -A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp --dport 5432 -j ACCEPT
<2> does it need udp?
<3> shouldn't
<2> k then will spend another 5 mins on google before giving up on pg again
<2> hmm seems like i need some postmaster thing running..
<2> nada under 5432 in netstat -plunta
<1> bleah
<2> hmmm mvcc in pg sounds cool
<2> anyone knows how exactly pg does hotbackups with mvcc, i mean how many snapshots can it afford to keep and what if while you make a backup a lot more transactions happen to a point where you have to get rid of things that are yet to be dumped for hotbackup...
<3> ask in #postgresql on freenode. :)
<1> this IS a php support channel ;)
<2> but ofcourse
<3> nobody cares if you're away.
Return to
#php
or
Go to some related
logs:
#linux
playing dota offline
#unixer
#kl
#india
#india
annatashia
#india
apr_recv: Connection reset by peer (104)
$SHELL -c $RUNWINE $FULL_COMMAND_LINE
|
|