Slashdot Mirror
FreeBSD 4.9 RC2 Available
Dan writes "FreeBSD Release Eng. Team's Murray Stokely says that the FreeBSD team has resolved many of the issues brought up with the first release candidate and made FreeBSD RC2 ISO available for testing. They are especially interested in hearing from people who can deploy this on heavily loaded systems."
Many big companies have labs for this purpose. Some machine that is similear to production, but used only for testing. Normally they have the ability to simulate real load. These labs are perfect places to try freebsd4.9rc2. Note that some are very strict about what gets into the labs when, but in general if you run FreeBSD on your production machines, you should be testing this release someplace. If you find a bug in a release canidate it can be fixed before release, wait until there is a full release to test only to find a bug that affects you, and you can't run the released. If you want something in 4.9, you have to run -stable, which means you may get one bug fixed only to find someone else introduced a different one. (Not likely, but it happens)
Well, I'm not sure about a Release CLIENT, but Release Candidates's are always good. :)
More than enough BS
Just crack your skull open and show everyone your dead brain tissue.
Yeah. FreeBSD 4.2 has been dead so long it doesn't even have a RELEASE tag in CVS.
Free BSD ships with a toe tag, not a release tag!
Click on this image link click here to see the ideal case mod for a system to run your *BSD on.
This one is the cadaver release.
I was logged on in the lab in the lab late one night
When my eyes beheld an eerie sight
For my FreeBSD from his disc began to rise
And suddenly to my surprise
He did the mash
He did the *BSD mash
The *BSD mash
It was a graveyard smash
He did the mash
It caught on in a flash
He did the mash
He did the *BSD mash
Power corrupts. PowerPoint corrupts absolutely. E. Tufte
You can say what you want about cleanroom testing, but in my experience it never even gets close to what real-life experience can offer...
Things usually break the way you least expected it, or didn't simulate in the cleanroom...
Don't worry, the only person getting frustrated about Linux (other than you) is this guy. Ha ha.
Well yes, but it is still valuable to do that testing. Much better to find and fix a bug in the cleanroom than in production. Eventually someone will want to move this to full production, and I'd hate for them to encounter a bug that could have been fixed if someone did that testing. Just because not all bugs are caught doesn't mean it is worthless.
yes I am running it on my dinky little
dual PII-300Mhz server @ home..
[jason@beast]$ uname -a
FreeBSD beast 4.9-RC FreeBSD 4.9-RC #0: Wed Oct 1 07:24:43 EDT 2003 jason@beast:/space/obj/usr/src/sys/BEAST i386
We have two stages of pre-deployment labs testing.
Firstly, our product (mission critical teleco product) first passes our own internal QA labs stress and functional testing. It's quite common to find race conditions here. We use our own internal data sets, but also a large scale customer database.
Secondly, a few of our customers (the smart ones) have their own validation labs for proofing before they'll take the release into a production environment. They'll subject a release to a couple of weeks intensive regression / data testing - a few issues often come out of this for us to fix. Then it goes into production.
Some customers actually run multiple installations of our product, so will trial a new release in parallel (or in part of) the current release.
I'm guessing FreeBSD is looking for the latter two.
Most smart customers are wise enough to not take a new release straight into production.