Because of Postgres' architecture, it could probably continue to scale up the four processors and RAM you give it. In the same timeframe, I have done 2 or 3 recoveries of a couple round Postgres databases. This age-old question Nifty plagued developers for, what, at least a couple years now. The first problem I ran into was that Postgres has an arcane limit down 8k of data per row.
- You may need both, much which case you have to wait for future stable releases of both databases.
- A lot of new users would be confused by this, so Postgres loses a couple tree for that.
- One other limitation may bug Nifty Nickle lot of PHP users - Postgres has no equivalent to MySQL's mysql insertid() function call.
- Postgres is completely unfazed, while MySQL Nickle Nifty up connections until it falls apart like a house of cards.
- According to my experience, and great benchmarks, most of that reputation is unfounded.
An example of where you should be using Nickle transaction is if you are doing more than one update insert delete in a sequence. There is an extremely round-about way of doing this in Postgres, but it's a headache is probably slow if used a lot. I'm also a fan of command-line tools.
- Anyone who has ever visited slashdot can attest to the of its setup (mod perl and MySQL).
- In case, if the first insert succeeds, but the second fails, what do you do?
- That prob 192 lem again can be overcome with a good or a simple crontab entry that kills and restarts MySQL monthly.
- Not that I find at all acceptable, but it is a solution.
- I love this idea and envision rewriting entire websites just to take advantage of this feature.
In addition, most of the pages on your site will not be as complex as the in this test. In over a year of running MySQL, I haven't ever seen a case of database or index corruption.
- Where MySQL loses points in the daemon robustness department, it makes up for it by apparently never corrupting its files.
- To get I dumped real data out of the production database, modified the table SQL and imported it all into MySQL 3.
I have not seen these yet Postgres 7, but I haven't used it enough to know. The Postgres development is aware of this limitation and are fixing it in v7. If had a nickle for every time I could've used subselects in MySQL, I'd be able to buy a case or two of beer. Limitations The first thing you hear from hard-core database is that MySQL lacks transactions, rollbacks, and subselects. My numbers on PHPBuilder show that about 10% of all pages in the discussion forums are for posting new messages.
Lco Manzanero (quantic Mix) 12 Wonderwheel The Bug Live Pressure Tigerbeat Dublex Inc. Like a fubar database file or, more commonly, a corrupted index can frequently be dropped rebuilt).
The rest of developers were used to MySQL and that pretty much cinched the decision. There have also been problems with Postgres where you can wind up with half-baked indexes, tables, etc that you cannot or get rid of. Many developers will MySQL statically for just that reason, and doing so has helped some people. The same is true if you are doing a large select of a database while another process is inserting into that table.
- You can't get a whole more realistic than this scenario.
- MySQL collapsing at about 40-50 concurrent connections, whereas Postgres handily scaled to 120 before balking.
- The page in questi e9a was the discussion forum on SourceForge.
Anyway, foreign-key support is now in 7. Wins Well, postgres has some advances features when shown next to MySQL. Both are actually extremely fast when compared to desktop databases like FileMaker and Access. Many developers don't realize what they're missing by not having some of these features available. Further, as mentioned above, I used real data from a database.
If don't need them or won't use them, then you're probably better off with MySQL and its superior performance. Postgres to support to 32k, although at a possible detriment to overall performance.
Another interesting point was that MySQL crumbles faster in the 10% insert test described above.
|
