Re: Problems with booting via ethernet

[magic@on-line.de]

On  3 Mar , Ken Yap wrote:
>>Hi !
>>
>>I have got a problem with booting via ethernet. I've compiled
>>linux-2.0.30 for Root on NFS and configuered my NFS-Server to serve
>>my diskless Clients. I've applied a Patch by Harald Milz so the kernel
>>generates a few devices before starting init. But I'm getting error
>>messages after the client mounts the nfs-root. My client get's its
>>bootimage via tftp. After the client has mounted its root filesystem
>>tftp makes the error msg:
>>
>>Mar  3 11:42:26 magic in.tftpd[650]: connect from 192.168.0.2
>>Mar  3 11:42:26 magic tftpd[651]: tftpd: trying to get file: /local/tftpboot/
>>iX-Terminal/romeo/bootImage 
>>Mar  3 11:42:37 magic tftpd[651]: tftpd: read: Connection refused
>>
>>My client is getting his bootImage, uncompress it and after it mounts
>>the nfs-root, my nfsd hangs. 
>>
>>What is going wrong ??? Can anybody help me solve this Problem ? 
>>
>>I tried booting remote with an Win95-bootimage and everythings wents
>>well. I got a dos prompt an no error msg. I treid the same with an old
>>linux kernel lx-1.2.12 from the iX-Terminal-Package but my client is an
>>old i386 with no math-copro. I've tried recompiling it but gcc 2.7.2.1
>>don't like the old source code. 
>>
>>Jan
>>magic@on-line.de
> 
> This is strange. You say the client gets its bootimage ok but then
> tftpd (or more like tcpwrapper) complains that it cannot give out the
no, I don't think that he means the bootImage because there is ~ 12 sec.
delay between thease two msg. and the kernel starts up decompression

> bootimage. So is romeo/bootImage the bootimage or just some other file?
romeo/bootImage is the bootImage for the client romeo

> Is there in your client startup scripts that tries to fetch the file
> again?
the client startup scripts have never started 

> You say your nfsd (on the server?) hangs. Don't you mean the client
> hangs because it cannot mount the nfs-root? What is the last message
> from the client?
"getting filehandler for Root by RPC from magc.mad.de"
VFS: mounted root (nfs filesystem)

2 second's after that the client begins to send permanently NFS lookup
requests. 

like this:

14:14:10.252752 romeo.mad.de.1ace767d > magic.mad.de.nfs: 112 lookup [|nfs]
14:14:10.252752 magic.mad.de.nfs > romeo.mad.de.1ace767d: reply ok 28 lookup [|nfs]
14:14:10.262752 romeo.mad.de.1ace767e > magic.mad.de.nfs: 108 lookup [|nfs]
14:14:10.262752 magic.mad.de.nfs > romeo.mad.de.1ace767e: reply ok 28 lookup [|nfs]
14:14:10.272752 romeo.mad.de.1ace767f > magic.mad.de.nfs: 112 lookup [|nfs]
14:14:10.272752 magic.mad.de.nfs > romeo.mad.de.1ace767f: reply ok 28 lookup [|nfs]
14:14:10.282752 romeo.mad.de.1ace7680 > magic.mad.de.nfs: 108 lookup [|nfs]
14:14:10.282752 magic.mad.de.nfs > romeo.mad.de.1ace7680: reply ok 28 lookup [|nfs]
14:14:10.292752 romeo.mad.de.1ace7681 > magic.mad.de.nfs: 112 lookup [|nfs]
14:14:10.292752 magic.mad.de.nfs > romeo.mad.de.1ace7681: reply ok 28 lookup [|nfs]
14:14:10.302752 romeo.mad.de.1ace7682 > magic.mad.de.nfs: 108 lookup [|nfs]
14:14:10.302752 magic.mad.de.nfs > romeo.mad.de.1ace7682: reply ok 28 lookup [|nfs]

ooops: 
I've just mentioned that the syslogd restarted on romeo:
Mar  3 14:08:31 romeo syslogd: restart 

This happens for the first time (and I didn't changed anything this
time ???? &-[])

PGP signature