FreeBSD 5.3-BETA4 Available

BrunoC writes "Once again, the FreeBSD Project presents yet another beta release of FreeBSD 5.3. FreeBSD 5.3 BETA 4 features major bugfixes for ATA, 4BSD is now the default scheduler and overall stability has greatly improved. BETA 5 should hit the streets next week and should be the last BETA and a Release Candidate is scheduled too. 5.3 should be around by October 3rd. ISO images are available for those who want to help the testing process." (Use a mirror.)

Odd...

[keiferb]

For a dead project, they sure release a lot of betas...

Long live BSD! =)

Re:Odd...

[ValiantSoul]

FreeBSD has never been dead. It isn't heard of as much as linux however it is used in many cases such as being the server that runs Yahoo.

Gentoo's portage is actually based on FreeBSD's ports system which continues to grow at a great rate.

If i was a mod I would mod you troll for not knowing anything about it before posting...

Re:Odd...

[keiferb]

If you did, I'd have to mod you unable to detect sarcasm.

Re:Odd...

[ValiantSoul]

My appologies, I typed it at 2:00am while I had a little too much to drink.

The next morning I was thinking about responding to myself however I did not

atapicam

[dar]

I notice that FreeBSD now has a scsi emulation layer for atapi devices. Very nice. Might be time to take another look at FreeBSD.

Re:atapicam

[Sicnarf]

is this a new feature? seems to me they've had it before.

Re:atapicam

[euclid manatee]

I'm not sure when it became available, but it's there in my 4.10 system. I use to to write DVD-Rs.

Re:atapicam

[phoenix_rizzen]

Now has?? It's been in there since the middle of the 4.x series.

Scheduler?

[TheRaven64]

4BSD is now the default scheduler

How bizarre. They only just switched the default scheduler from 4BSD to ULE. Now they've swapped it back. Anyone know why?

Re:Scheduler?

[beholder77]

I think they said in the release that it had to do with stability issues with some optional kernel settings.

Dunno about the rest of you guys, but I find the 4BSD scheduler smoother on my single CPU boxes. Then again I have crappy hardware, so who knows why that would be :)

Re:Scheduler?

[anholt]

Part of it was that there were problems with ULE and PREEMPTION, which we really hope to have working for the release if possible. But it was also not performing well for some specific tests that are being done (supersmack on mysql, among others). That combined with a few little bugs (load avg reporting) and the author not really being available to work on it, meant that it was decided to move back to 4BSD for the release (and possibly -STABLE), while hopefully getting ULE going well on -CURRENT. ULE has its own set of features -- better interactivity on heavily loaded machines (so nice for your desktop while compiling!) and, as I remember, better algorithms for machines with tons of processes, but its youth menas it's not well-tuned yet for general server use, which is what FreeBSD focuses on.

Re:Scheduler?

[drmerope]

The author of ULE has been out of touch now for sometime. ULE had known issues at the time it was made default in -current, but it was hoped that visibility would generate the reports necessary to stablize it. Unfortunately, Jeff hasn't seemed to have the time to fix or even examine the issues uncovered.

The recent preemption changes drove ULE from being merely suboptimal (inefficient) in some situations to being unstable.

Switching

[rawg]

I switched over to FreeBSD a while ago because Debian was so out of date. I fell in love with FreeBSD because it's so easy to maintain. And Ports are awesome! I can run a stable OS with the most up-to-date (stable) software. Before I had to run unstable Debian to get the latest PHP/Postgres to install, unless I compile it myself.

Re:Switching

[rawg]

I think your wrong. I was bit too many times when I did a apt-get update/upgrade and everything broke. The last time it happened I switched to FreeBSD. I have never, not even once, had a problem with stability in FreeBSD. I have the most updated stable software running, and the OS is stable. I'm talking about FreeBSD 4.10, Postgres 7.4, PHP 5, ... They are all labeled as Stable. And I have not had a single crash, seg fault, or loss of data.

FreeBSD just works as advertised, "Stable".

With Debian, I w

Re:I read this recently, how accurate is it?

[ValiantSoul]

Enabling power management in the kernel would help. You are most likely comparingn linux w/power management vs FreeBSD w/o

Re:Misplaced effort...

[R.Caley]

And yet they still haven't fixed [the block count thing]

Well, neither have you, and you have all those numbers in your name to prove how cool you are.

Re:Misplaced effort...

[SillyNickName4me]

From your entire bug report, it seems pretty obvious that something is rather wrong with the ata controller and/or bios on your hardware.

I happen to have had 3 of the same Maxtor drives as you have, and the last surviving one is currently primary master in my router. (hint, replace them, I have had 2 of them give up within short time of eachother, and the 3rd one seems to be getting close to giving up)

All 3 have always worked with FreeBSD, Linux and Windows, without any confusion with regards to its CHS settings.

I did however follow the rather strong suggestion that my BIOS gives me to run the bios auto detect and select 'normal' mode for a drive on which I am going to use a unix like system. This results in a user configured drive (for as far as the standard cmos settings go)

Oh, and I also followed the recomendations to have the disk as primary, and a cd drive as slave instead of having it as primary like you have.

That it is not recognized that way during sysinstall seems rather suspicious to me.

Maybe FreeBSD needs some patches for detectign this broken hardware configuration, just as it did get a patch for dealing with the rather broken bios of the asus p2b-ds that I happen to use (apic is broken, resulting in an interupt storm when doing things according to the official standard)

If you think this problem affects more then your very specific case, I suggest talking about it on the mailinglists for -current, and try to be helpfull in getting a workaround, ie, that means accepting that it is in fact a problem of your hardware.

All that said, your comment regarding fdisk seems to be correct, and it should accept the alternative geometry if those fit within the physical number of sectors that the drive has.

Re:Misplaced effort...

[hugo_pt]

Actually, this seems like a real problem as on at least two workstations I've had fdisk insisting the drive geometry was wrong. Specifiying the correct vallues to fdisk didn't help as it insisted with the old values.. But ultimately I could install the OS just fine even with the geometry set wrongly and it reports the partition sizes well.

I've been happy with Beta3

[Moloch666]

I switched over my 5.2.1 personal server to 5.3-Beta3. Mostly for the NDIS driver so I can get my atheros wireless card to work in turbo mode. It runs great. Their ath drivers are much more mature than linux's madwifi they still can't do turbo, all though the man file said otherwise. I'm guessing it's an issue with my card(DWL-G520). I now have my 2 linux computers and this one using windows drivers using ndis wrappers. My wireless speeds are now up to acceptable.

I suppose I'll give this one ago, although by the time I get around to it, I'll probably be cvs up'ing beta5.