[ENBD] More nbd-2.4.1 questions on kernel 2.2.18

Peter T. Breuer ptb@it.uc3m.es
Sat, 10 Mar 2001 02:17:08 +0100 (MET)


"A month of sundays ago Kai Chen wrote:"
> >Most certainly you are mistaken. That is a prerequisite, not a
> >symptom.
> 
> I did do "make config all."

Did you remove config.cache first?

> And during the compilation, I did see things like:
> 
> gcc -O2 -Wall -D_LARGEFILE64_SOURCE -I/home/kylechen/nbd-2.4.21/tmp 
> -I/home/kylechen/nbd-2.4.21/linux/include  -DDEBUG=0 -o nbd-server.o -c 
> /home/kylechen/nbd-2.4.21/nbd/nbd-server.c

This doesn't matter. It's the config.h entries that control which seek
function is used. The options here merely ensure that what you want to
happen _can_ happen, not that it will.

> So I think that part should be fine.

Not so. Look at config.h.

> What I was wondering about was if I should have done something to, e.g., the 
> kernel to support these "large files?"

No. The problem is in userspace .. mind you, if you _really_ mean
files, and not partitions, then yes, you can't use a file larger thab
2GB on an extfs partition without kernel patches. At least that used to
be so.

> I didn't see "config.cache," but I did "make clean" and "make ditclean," so 
> I think the configuration cache was removed.

OK.

> >I suspect that you really haven't run make config. At least look in 
> >config.h and see if the configured values are appropriate for your system.
> 
> I'm attaching my config.h file at the end of this message.

I don't see it. Perhaps because I'm replying. Please use ordinary
quotes - I can't see an attachment until I finish replying to you!

> >>nbd-client-netserver: client (0) short read from net to buffer  offset 0, 
> >>wanted 4096 got -110
> >
> >That's saying it couldn't even succeed in negotiating the handshake.
> >The mkfs would never have run.
> 
> Unfortunately "mkfs" did run successfully even with the above error 
> messages.

It's not possible. It must have been on an earlier connection.

> I'm running Redhat 6.2 with kernel 2.2.18 everywhere.  There should be no 
> mismatch...
> 
> I use PIII 800 128M RAM with both SCSI and Fibre Channel Drives as my 
> server.  The clients are PIII's or Celeron's.

Looks fairly normal.

Peter