Last week, I posted a question about some problems
I was having with RFS. Here is a portion of the
original message:
>It seems that 3b2 cannot access files with inumbers greater
>than 65535. At first, this would seem to be a 3b2 problem,
>but the Sun is returning the error status. I've been able
>to confirm this. I've found an ls -l command, as it stats
>every file, returns errno 24 (EMFILE) to the 3b2 for every
>file with an inumber greater than the magic 65535. (well,
>I didn't check 65534 and 65536 to find the exact cutoff, but
>61000 works and 70000 doesn't.)
>
>If you've experienced this, please drop me a line.
>
I received a number of responses, the gist of which is
succinctly stated by Hal Stern, so his is the only response
I'll reproduce here...
-----------------------------------------------------------------
this is a bug or misfeature in Sun's RFS implementation;
it's documented in the Read This First for 4.1.1(b).
sun's rfs uses 16-bit inode numbers, and returns EMFILE
if you try to access an inode > 64k.
so you need to export only 131M or smaller filesystems
(so you have < 64k inodes). or else rebuild the filesystem
with a much lower inode/dblock density
--hal
-----------------------------------------------------------------
Thanks to:
uunet!ikarus.ts.go.dlr.de!eckhard (Eckhard Rueggeberg) [twice]
Mike Raffety <uunet!sbcoc.com!miker>
uunet!thor.cs.wayne.edu!jjb
uunet!lgc.com!keith
uunet!cis.ksu.edu!jfy (Joseph F. Young)
uunet!sunne.East.Sun.COM!stern (Hal Stern - NE Area Systems Engineer)
uunet!cmcl2.NYU.EDU!tester (thx1138@nyu.edu)
for the responses.
I've not yet decided how to work around this. I rather wish I posessed
the Read Me first so I knew this before I set all this equipment up,
formatted my disks, etc. I could split up the file system and export
all of them. I could put the info on the 3b2 and export it back to
the Sun, which is actually how it is set up now. But this has its
problems too, which I won't go into here now. Additionally, I was
planning on freeing up the 3b2 partition so I could use it for something
else.
Thanks again for the help..
Gregg Siegfried
grs@sacs.com
grs@ihspot.att.com
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:35 CDT