FreeLists / openbeos / [openbeos] Re: - actually: [off-topic] and [selfish]

["Axel Dörfler" <axeld@xxxxxxxxxxxxxxxx>]

Jared <jared@xxxxxx> wrote:
> i hate to do this - but i'm posting a self-important SOS "please 
> help" 
> sort of thing today.  *sigh*  and i was always secretly proud of 
> being 
> idealistic and keeping my posts on-topic (few as they may be)

This kind of SOS is perfectly fine with me - dunno about others, though 
:-)

> so i've booted off the R5 pro cd and ran chkbfs -f
> it 'fixed' several hundred thousand lost clusters (i forget what 
> nomenclature was used there...but lost clusters they were)

It's usually block_runs or blocks.

> it now runs "clean" through chkbfs, but nonetheless i can see this 
> error 
> above the "no errors" satement:
> Failed to load (NULL)/beos (ino # 1049569)
> Files processed: 11650read_vnode : (bi @ 0x805ef008 104972) 
> betree==NULL??!? (thid 261)

I can only guess what this means: the inode which is a directory (bi ==
 bfs_inode) has an invalid B+tree. The tree is how a directory stores 
its contents in BFS.
If that error didn't exist at first, but only after the first "chkbfs" 
run, it is likely that chkbfs has removed the B+tree from the inode 
without updating the inode itself.
You can use my BFS-tools to find out which directory is affected (using 
bfsinfo). If it's not an important directory, you can then just delete 
it - if that doesn't work (likely), then use the "forcerm" command that 
comes with BeOS for those cases (they knew their bugs).
"chkbfs" should then run without showing this error any longer.

> so, a point of clarification, i'm boned, right?  does "bi" mean i 
> have a 
> bad spot on the drive or something?  Axel? (wink wink - nudge nudge -
>  
> say no more say no more) i remember you boasting bout your bfs 
> tools.... 
> if i get my hands on another blank drive and install BeOS on it (or 
> Zeta 
> when it arrives) you think maybe i'll get back my old install?  or 
> should i just break down and sob now?
> 
> anyways, there are a number of other chkbfs erors reported concerning 
> data in other places, but this seemed to be the only one that would 
> be 
> critical to loading the OS (yup, can't find an OS to load).

Please copy the contents of the chkbfs output and send them to me, if 
you want to know for sure.
If the file system can still be mounted, there are probably only two 
reasons why it couldn't boot:
1) the boot block is corrupt - use the "makebootable" command to 
rewrite it
2) the loader or the kernel is corrupt, replace them from your BeOS CD.

2) is less likely, I would guess that your system will boot normally 
after you have used "makebootable" on the partition.

Bye,
   Axel.