NetBSD-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: TFTP spawn issue



Hmm. My bad then. I thought I had run tftpd from the command line in the past, but I must not have, since this would suggest it needs inetd to change stdin (I assume) to the socket.

My only remaining suggestions right now are to put -d in the inetd.conf file, and check the logs. And possibly also run tcpdump and check what is going on. Or else start digging in the code.

        Johnny

On 2012-03-27 21:17, IT geek 31 wrote:
Yeah I tried that and it mugged me off :-S

I tried running:

/usr/libexec/tftpd -l -d -s /tftpboot/

And all I get in /var/log/messages is:

Mar 27 21:14:30 qube1 tftpd[430]: recvfrom: Socket operation on non-socket

I also don't know why it is retrying as it is definitely receiving the
files... the client (in this the ESXi 5 installer) helpfully outputs
the files it has loaded as it progresses....

And now I am stuck.


-Mark


On 27 March 2012 17:30, Johnny Billquist<bqt%softjar.se@localhost>  wrote:
man tftpd ? :-)

In short, you just run it. Preferrably with the same switches as specified
in inetd.conf

You probably also want to add the "-d" switch...

As for the 7 requests, it's probably retrying as it isn't getting responses.

The interesting part is probably what you see in the syslog.
I'd need to dig in a bit further to know more in detail, but I don't have
time for all that fun right now. :-)

        Johnny

On 2012-03-27 17.12, IT geek 31 wrote:

Hi all,

Thanks for your advice.  I'm inclined to agree with Johnny - something
isn't right and I don't think increasing the spawn rate is solving the
issue.

I have turned on debugging for inetd, although I'm no clearer.

When the client boots, it appears as if it's asking for each file 7
times.  /var/log/messages shows this:

<snip>
Mar 27 17:06:10 qube1 tftpd[1782]: esxi.mdb.eu.com: read request for
//esxi50/ehci-ehc.v00: success
Mar 27 17:06:10 qube1 tftpd[1570]: esxi.mdb.eu.com: read request for
//esxi50/ehci-ehc.v00: success
Mar 27 17:06:10 qube1 tftpd[1688]: esxi.mdb.eu.com: read request for
//esxi50/ehci-ehc.v00: success
Mar 27 17:06:10 qube1 tftpd[1724]: esxi.mdb.eu.com: read request for
//esxi50/ehci-ehc.v00: success
Mar 27 17:06:10 qube1 tftpd[1759]: esxi.mdb.eu.com: read request for
//esxi50/ehci-ehc.v00: success
Mar 27 17:06:10 qube1 tftpd[1598]: esxi.mdb.eu.com: read request for
//esxi50/ehci-ehc.v00: success
Mar 27 17:06:11 qube1 tftpd[1186]: esxi.mdb.eu.com: read request for
//esxi50/ehci-ehc.v00: success
</snip>

The console is outputting this:

<snip>
2245 execl /usr/libexec/tftpd
2245 reaped, status 0
restored tftp, fd 5
someone wants tftp
</snip>

In the end the client will refuse to boot, and ask to be rebooted.
/var/log/messages shows this error:

Mar 27 17:07:05 qube1 tftpd[1603]: Received *code 768* in sendfile
Mar 27 17:07:25 qube1 last message repeated 4 times

Any idea what that means?

I have checked the md5sum of the ISO and unpacked it again - the boot
files seem good.

Johnny - how do I run TFTP by hand from the command line?

Again, thanks to all for your input,


-Mark


On 26 March 2012 21:47, Johnny Billquist<bqt%softjar.se@localhost>    wrote:

I think that is bad advice. If the max spawn rate is exceeded, there is
something wrong. Increasing the max spawn rate will not solve that, but
can cause greater problems. However, I suspect the end result will just
be the same, but it will have forked more times before disabling the
service.

What I'd do is I'd disable the service in inetd.conf, and run it by hand
on the command line, with debugging enabled. That would probably give a
hint on what the problem is.

        Johnny

On 2012-03-26 12:56, MASUDA Hideo wrote:

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Mark,

   Try this configuration.

/etc/inetd.conf ---
tftp dgram udp wait:400 root /usr/libexec/tftpd       tftpd -l -s
/tftpboot
#                  ^^^^<- increase max spawn rate
#                          from 40 (default) to 400 (example).

   See Also inetd.conf(5).

Cheers.
- --
MASUDA Hideo (h-masuda%ootani.nagata.kobe.jp@localhost)
FingerPrint = C532 9F5B 4D7D DA76 8BFF  CAA3 C073 5C9E 10FF 98EF

2012/03/24 00:37, IT geek 31 wrote:

I am trying to PXE boot an OS installation media using both PXE and
TFTP.

I have configured DHCP, the PXE menu file and the appropriate file
structure without issue.

However, as my client boots, on the NetBSD server I receive the
following error in /var/log/messages:

            Mar 23 16:35:41 qube1 inetd[596]: tftp/udp max spawn rate
(40 in 60 seconds) exceeded; service not started

At this point it appears the TFTP daemon crashes and the OS
installation on the client fails.

I have tried changing the following line to nowait in /etc/inetd.conf,
but after a reboot, TFTP refuses to start:

            tftp            dgram   udp     wait    root
/usr/libexec/tftpd      tftpd -l -s /tftpboot

Anyone have any idea what I need to do to get it working?  I'm running
NetBSD/cobalt 5.1.

Many thanks,


-Mark

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Vine - http://enigmail.mozdev.org/

iQEcBAEBAgAGBQJPcEtkAAoJEMBzXJ4Q/5jvQPQH/3eWdf+MD58DD5TnwnaGqzfT
9qSk1uPl2mzxI4uIJ4BFfLfkNAkHbPw5iDpawy85mTRG06KiXyHp/6iljvNH8+gy
ELo0JpOz3fz0MBnxSWIbR9RLZluqDbk6JWJidX13j2dO+ZoQN6v7nxIW/b//RsA8
tCoVTLiXwC59j5Mz5tmFByLoWLN7cPpCBq830LG7neCd6QGhs6N+SEVgUoBsKMiR
dQ5jCHwUcmNxcfNYnMTGVdAylwQVvx1pHjhaMcmywM3E0UaoxMbBco4xpG0nLHqQ
0JhjjYBvv8SD/dwsBK+SmJGsZF1KRX2+xsy+J9OQHOMr3XWFFDgyH/SCqJ+AE5Y=
=R5jH
-----END PGP SIGNATURE-----



--
Johnny Billquist                  || "I'm on a bus
                                  ||  on a psychedelic trip
email: bqt%softjar.se@localhost             ||  Reading murder books
pdp is alive!                     ||  tryin' to stay hip" - B. Idol




--
Johnny Billquist                  || "I'm on a bus
                                  ||  on a psychedelic trip
email: bqt%softjar.se@localhost             ||  Reading murder books
pdp is alive!                     ||  tryin' to stay hip" - B. Idol


Home | Main Index | Thread Index | Old Index