PFS3 Does Not Install - Error

[Piru]

It appears that the release archive is missing the DirectSCSI version of the filesystem for 68000/68010 (PFS3/data/#?/fs/#?PFS3ds). If you install on 68020, 68030, 68040 or 68060 system it should work just fine.

Assuming you're running on 68000 or 68010 and want to install you can select the expert mode where you get to choose the CPU type (choosing for example 68020 will bypass the problem). After the installer has completed you can manually copy over the existing 68000/68010 filesystem versions.

Of course this doesn't help the fact that 68000/68010 directscsi version is missing.

[Piru]

I switched from intermediate install to expert. This got me around the problem and got to choose the 030 version.

Then hit another missing file:

You really need multiuser support? If not, don't select them at install ("Multiuser version" and "DirectScsi Multiuser version"). If you select multiuser versions the installer attempts to patch the *existing* multiuser installation.

[Piru]

I did get it installed and all my enthusiasm for it is now extinguished.

Well too bad. The filesystem really is the best there is.

[Piru]

Not true... SFS is much better

• PFS3 is faster than SFS (easy to benchmark and verify as a fact)
• PFS3 fragments less than SFS over time (easy to verify as a fact)
• PFS3 has a repair tool, SFS doesn't (except MorphOS SFSDoctor)
• PFS3 is more reliable than SFS (somewhat subjective and harder to verify)

What SFS has going for it is the support for larger partitions than PFS3 (SFS: 128GB PFS3: 107GB)

[Piru]

So again we're in agreement that PFS3 is superior. Great!

[Piru]

Think you need new goggles...

Nope, I can see perfectly fine that you provided no evidence whatsoever that would change my assessment of the facts. Claims 1, 2 have been proven to be correct numerous times in the past, by numerous different people. Claim 3 is a plain cold fact. But as we well know you won't bother, instead resorting to just making some seemingly witty comments. Some claim we need village idiots, I'd say we can do just fine without.

Should've gone to SpecSavers...

(You probably wont get that one, it's a British TV ad...)

Oh but I do. We do have specsavers. No, I don't shop my glasses from there.

[Piru]

I would say that if PFS was updated to support bigger Hard Disks, than it would be no contest as to what to choose.

I guess you mean larger partitions, PFS3 already supports as large HDDs as SFS.

@Piru

Was PFS ever successfully ported to MOS natively?

It works. I've ported the filesystem, pfsdoctor, diskvalid and all the related commands. I'd need to do some more testing, but unfortunately my Peg2 started acting up lately. I've moved my devenv to my 2nd Mac Mini now, but I haven't had much time to work on it lately.

[Piru]

Great, so PFS3 will be on the 2.8 CD?

It will be inside the bootimg, but whether it'll make it to 2.8 remains to be seen. I'll try my best.

And whats up with the Pegasos, I've been hearing people are having a ton of issues on them nowadays.

I guess it's just the age. This particular board is from the first batch so it's quite old already. It works fine for some time but then the IDE bus appears to drop in/out constantly. As if the HDDs would be running out of power (but they're not, I tried with another PSU and it's the same). As I said I've already migrated to the Mac mini so it's no real issue.

[Piru]

How big can a PFS3 partition be?

about 106GB

[Piru]

When your work on PFS3 is ready, maybe you can look into addressing those limitations that still exists, and call it "PFS4" of such, so "backwards compatibility" to old PFS3 disks can be ignored with good conscience?

It looks like that'll have to be the way. Some of the ways to overcome this limitation have been discussed here: http://sourceforge.net/tracker/?func=detail&aid=3285711&group_id=532591&atid=2163218

I really don't like these "lets double the capacity" tricks much, as they don't raise the limit enough really. I think another level of indirection is the way to go.