Subject: ufs.....
To: None <>
From: Reinoud Koornstra <>
List: netbsd-users
Date: 09/26/1999 03:03:23
Hi All,

First of all, i have at my second disk freebsd divided over two slices and
i have one last slice on that disk for netbsd.
Now... i made the slices for freebsd with the freebsd fisk.
When i let linux look at those slices, then it says that those slices dont
start at the beginning of a cylinder and not ending at one or soimething
like it. ...So it reports errors to that slices.
Now i made with the linux fdisk a slice for netbsd and changed it
while installing netbsd. Then i made labels in the slice.
When i sometimes boot freebsd, it said that the labels in the netbsd slice
are not completly within the slice and freebsd truncates them.
So i see, that linux, freebsd and netbsd handle the disks quite
differently. Which one is the best in handling the disks.
Cause all reports errors to one another etc etc....
Who is right?

Secondly, imagine i have a disk of 150 gig. And i want to put netbsd on
it. Is that possible without problems?
What if i have a directory in a partition in a slice wich contains
1.000.000 files in it...
I believe that the ufs cannot handle this, cause the inodes are not
enough. The xfs can handle this without a problem cause it doenst work
that way with inodes i believe. Also, because of the logging of the file
system it can nearly always be recovered. Also when the partition are
gone. Imagine of course that i have the loggin of the xfs on a different
disk. Now i heared that the ufs works with the synchronisation of
metadata. What does that mean exactly? Can ufs also always be recovered?
Also when even the slices are gone?
When you turn of the power.... the ufs needs a little time to recover
itself. The xfs doenst need anytime to recover because of the logging.
Well.... reading about the xfs i get the impression that the xfs is a
little better then the ufs.
Furthermore..... i cannot read information about the ufs anywhere about
the recovering possibilities. About the synchronisation etc etc.
I searched the net. Anybody knows where i can find things about this?
Is there also a logging fs for netbsd?
There is one but......... for large partitions the logging fs can become
very slow i believe. But... another nice thing of an lfs is that when you
overwritten a directory, you can recover it because of the logging.
I want to learn much more of the possibilities of the ufs and
impossiblities. What are the plans for ufs in the future etc.
Is there something gonna be done about the inodes and the way ufs handles
them. Cause how can the ufs ever handle dirs with many files and such and
being still able to use the resty of the disk space?
Right now, i tend to think that the xfs is a little bit better (what is
better?????) the the ufs because of what i mentioned above. Plz forgive my
ignorance if i am completly out of line here, but right now, i think so.
Hope to hear something.