Re: [9fans] kfs question

Re: [9fans] kfs question

[rsc]

After you save what you care about, you can drop
the whole /mail/box hierarchy by doing

	disk/kfscmd 'clri /mail/box'

which will remove the /mail/box directory entry pointer
from /mail and do nothing else, leaking all the storage
associated with /mail/box.  Then

	disk/kfscmd 'check f'

will rebuild the free list, reclaiming the storage.

Russ

Re: [9fans] kfs question

[Micah Stetson]

> After you save what you care about, you can drop
> the whole /mail/box hierarchy by doing
>
> 	disk/kfscmd 'clri /mail/box'
  ...
> 	disk/kfscmd 'check f'

Thanks a lot; that worked.  I'm glad I didn't have any files
that I cared about in there as anything like 'lc /mail/box'
would cause kfs to die immediately.

Micah

[9fans] kfs question

[Micah Stetson]

I was trying to build a distribution floppy image on my
laptop (486-75, 32MB, ~500MB HD, I know, old, old, old),
but kfs died every time I attempted it.  A couple of times,
it just paniced, once it said it ran out of swap space.
While I'd be interested to know what caused this, I have
a more immediate problem, now.  Since kfs died, I couldn't
halt the filesystem, so after a couple of crashes of this
sort, I ended up with some corruption.  Now whenever I
run kfscmd check, I get a bunch of complaints just like
this one:

/mail/box/->name ('€\x0e\x02'€\x0e\x02'€\x0e\x02'€\x0e\x02'€\x0e\x02'€\x0e\x02'€\x0e) not terminated

since I don't imagine many mail gateways will like that, here it is
piped to xd -c:

0000000   /  m  a  i  l  /  b  o  x  /  -  >  n  a  m  e
0000010      (  ' c2 80 0e 02  ' c2 80 0e 02  ' c2 80 0e
0000020  02  ' c2 80 0e 02  ' c2 80 0e 02  ' c2 80 0e 02
0000030   ' c2 80 0e  )     n  o  t     t  e  r  m  i  n
0000040   a  t  e  d \n
0000045 

My question is, is there any way of fixing this, I've tried
disk/kfscmd 'check rfdtw' to no avail and disk/kfscmd
'remove /mail/box/all that junk' says that there's
an overlong path element.  Doing disk/kfsmcd 'remove
/mail/box' doesn't do anything presumably because the
directory is not empty.

As a side note, the synopsis of possible arguments to
check in kfscmd(8) is different from the list that is
documented below the synopsis.  The synopsis has no 'c'
option and the description has no 'R' option.  Looking at
the source, there doesn't seem to be an R option, although
there is an undocumented 'q'.

Thanks in advance,

Micah