Pure BS. There is no way in hell postgres can come close to 35k inserts/s. To get 35k inserts/s you would need a high end cluster system... running something other than postgres.
Why? InnoDB was already GPL'd. Why not continue to use and develop it?
Because mysql sells a private version of their DB server. They can't do this with GPL'd code in it. They had a separate (non-GPL) license aggrement from Innobase Oy to use the InnoDB engine.
My company has a semi-realtime application that needs to insert ~200-2000 rows/sec constantly. It isn't true realtime because the db can be shutoff, rebooted, etc and the inserts will queue waiting for the server to come back online.
Approx 1 year ago, we were doing some enhancements on the application and I tried replacing the mysql backend with postgresql. We needed plenty of expert help tweaking postgres to get it to a point where it could keep up with the inserts, however we could not run the vacuum job. There was no way to measure the exact performance difference between mysql and postgres, but I estimate mysql (innodb) to be able to handle 5x the load of postgres.
Why would I want to use a DB server that can only handle 20% the load of a competitor?
Only DB1 is open in a way that is useful to commercial software vendors (ie the folks at Mysql). DB1 is BSD licensed while DB[2-4] are licensed under the Sleepycat license which is only free for GPL software (basically).
While mysql (the db) is GPL'd, it is also proprietary. Mysql (the commercial entity) has to abide by the Sleepycat license if they distribute private versions of the server, so Oracle purchasing Sleepycat has a chilling effect on Mysql (the commercial entity) and the DBD backend.
CXFS sounds very interesting, but it has some unfriendly dependencies. Think GFS that Redhat recently bought but with usable performance. With a SAN you can attach up to 256 different machines mounting the same filesystem.
The downside is you need one machine running IRIX controlling the cluster. All the clients can be running linux if you wish. If you have existing XFS volumes they wont even need to be reformatted to run CXFS.
Maybe it will open source now and the IRIX dependency will go away?
Re:FreeBSD is cool like all other OS's
on
Why FreeBSD
·
· Score: -1
Actually its linux which has the fsck'd threads. Every thread is actually a process and has a pid??? NPTL seems to be going in the right direction though.
The problem with ipv6 is that there is too much address space. IP addresses are worth anything in ipv6 which is actually a huge problem.
ipv4 addresses are nowhere near as common as ipv6 and look what happens to those addresses, they get permanantly added to various blacklists, various routers will drop all packets from certain addresses, etc.
ipv6 with its near infinite address space makes ip addresses close to worthless which is a huge problem. ipv6 will never take off.
Couldn't be further from the truth. Last week I was playing around with Qemu and a friend already had a Windows 95 image ready to go. So I was just using Windows 95 last week, and I can say that the only real difference is tabbed browsing (which rocks).
I don't know where you'd get the idea that I have a lack of knowledge? I work exclusivly with *nix servers at work doing both programming and system management. And to top it off I use linux and KDE on my desktop. So actually I've got a wealth of knowledge about the subject, thanks very much!
Basically you are the typical fanboi, anytime someone posts something contrary to your one true FOSS way, you call them naive or incompetent.
Sorry to piss in your cornflakes, but you're the one with the "lack of knowledge".
Slashdot Mirror
Pure BS. There is no way in hell postgres can come close to 35k inserts/s. To get 35k inserts/s you would need a high end cluster system ... running something other than postgres.
/. mod parent UP!!!
Anyway this is
postgres has no clustering support. It has limited replication support from a project called slony (http://gborg.postgresql.org/project/slony1/projdi splay.php).
The mysql clustering is pretty much bunk anyways. All of the DBs need to live in RAM. What happens if you have a catastrophic UPS failure?
Because mysql sells a private version of their DB server. They can't do this with GPL'd code in it. They had a separate (non-GPL) license aggrement from Innobase Oy to use the InnoDB engine.
http://www.jcb-sc.com/qmail/guninski.html0 _ tut.pl?tutorial_name=Qmail_vulnerabilities.html&fa ct_color=doc&tag=
http://secunia.com/advisories/10649/
http://secunia.com/advisories/15533/
http://www.frsirt.com/english/advisories/2005/049
http://www.frsirt.com/english/product/3207
http://www.saintcorporation.com/cgi-bin/demo_full
My company has a semi-realtime application that needs to insert ~200-2000 rows/sec constantly. It isn't true realtime because the db can be shutoff, rebooted, etc and the inserts will queue waiting for the server to come back online.
Approx 1 year ago, we were doing some enhancements on the application and I tried replacing the mysql backend with postgresql. We needed plenty of expert help tweaking postgres to get it to a point where it could keep up with the inserts, however we could not run the vacuum job. There was no way to measure the exact performance difference between mysql and postgres, but I estimate mysql (innodb) to be able to handle 5x the load of postgres.
Why would I want to use a DB server that can only handle 20% the load of a competitor?
Only DB1 is open in a way that is useful to commercial software vendors (ie the folks at Mysql). DB1 is BSD licensed while DB[2-4] are licensed under the Sleepycat license which is only free for GPL software (basically).
While mysql (the db) is GPL'd, it is also proprietary. Mysql (the commercial entity) has to abide by the Sleepycat license if they distribute private versions of the server, so Oracle purchasing Sleepycat has a chilling effect on Mysql (the commercial entity) and the DBD backend.
If I only had real time firefox I could've made this first post!
CXFS sounds very interesting, but it has some unfriendly dependencies. Think GFS that Redhat recently bought but with usable performance. With a SAN you can attach up to 256 different machines mounting the same filesystem.
The downside is you need one machine running IRIX controlling the cluster. All the clients can be running linux if you wish. If you have existing XFS volumes they wont even need to be reformatted to run CXFS.
Maybe it will open source now and the IRIX dependency will go away?
A free as in speech access to files for Apple owners? Put down the crack pipe!
That is the way UDP works.
Nobody saw this coming ...
Actually its linux which has the fsck'd threads. Every thread is actually a process and has a pid??? NPTL seems to be going in the right direction though.
Pretty much every post I've every made here has been moderated down by tards.
slashtards they be!
The problem with ipv6 is that there is too much address space. IP addresses are worth anything in ipv6 which is actually a huge problem.
ipv4 addresses are nowhere near as common as ipv6 and look what happens to those addresses, they get permanantly added to various blacklists, various routers will drop all packets from certain addresses, etc.
ipv6 with its near infinite address space makes ip addresses close to worthless which is a huge problem. ipv6 will never take off.
The nextgen packaging system is already in Gentoo.
emerge -- use it, live it, love it!
As I mentioned in another post, I just "went back" last week and used Windows 95. There was no rude awakening.
... the damn thing kept locking up! But the subject isn't about OS robustness.
Certainly it isn't fit to use for a day or two though
How many incremental improvements can you actually name?
- tabbed browsing
- ???
I don't know where you'd get the idea that I have a lack of knowledge? I work exclusivly with *nix servers at work doing both programming and system management. And to top it off I use linux and KDE on my desktop. So actually I've got a wealth of knowledge about the subject, thanks very much!
Basically you are the typical fanboi, anytime someone posts something contrary to your one true FOSS way, you call them naive or incompetent.
Sorry to piss in your cornflakes, but you're the one with the "lack of knowledge".
Cheers!
I don't know if anyone's been paying attention, but the UI has basically stagnated since the mid nineties.
Is XP/KDE/Gnome really all that different from Windows 95 and Netscape 3.x?
And to answer my own question, no it is not.
Certainly no UI innovations are coming from FOSS -- all they can do is copy Micro$oft and Apple.
What fucking retard modded this to a troll?
http://c2.com/cgi/wiki?XmlSucks
nt = no text
XML sucks ... too verbose for humans and too ambiguous for machines.
One day we'll look back and laugh!
What else could it be? Perhaps you were thinking that you could emulate a beowulf cluster on a single P4?
Note that this accelerator only works on x86 emulating x86 so if you're thinking about z80 and G5 then you need to RTFA ...
There is no way in hell you can hash a 700MB file in 0.01 seconds.