tftp boot issues...

[Chris]

Strange fault with tftp on Solaris 10. Originally
set up for an install server a couple of years
ago, which did work. Added case for a Sun 4C
machine, Solaris 2.6 install.

The files and setup seem ok, and the tftp.udp6 svc
starts up and comes online. As soon as the client is
pointed at it, it transitions to maintenance mode.

Here's some log info

* inetadm says:

enabled maintenance svc:/network/tftp/udp6:default

* /var/adm/messages says:

"inetd[311]: [ID 702911 daemon.error] The inetd_start
method of instance svc:/network/tftp/udp6:default
isn't a regular file".

?????

tftp boot dir is at: /export/tftpboot and this
is known working for the original clients.

Startup line in /etc/inet/inetd.conf has been
updated to reflect that.

Ip address for client is at 192.9.200.21

* /export/tftpboot line for host says:

C009C815 -> inetboot.SUN.4-75-solaris.26

points to:

173280 Jul 16 1997 inetboot.SUN.4-75-solaris.26

* /etc/hosts entry:

192.9.200.21 ss2sol26

* /etc/ethers entry:

08:00:20:c0:11:07 ss2sol26

* tcpdump ether host 08:00:20:c0:11:07 (trimmed date & time)

rarp who-is ss2sol26 tell ss2sol26
rarp reply ss2sol26 at ss2sol26
IP ss2sol26.3549 > darkstar.tftp: 23 RRQ "C009C815.SUN4C" octet
IP ss2sol26.3549 > 0.0.0.0.tftp: 23 RRQ "C009C815.SUN4C" octet
IP ss2sol26.3549 > 0.0.0.0.tftp: 23 RRQ "C009C815.SUN4C" octet
IP ss2sol26.3549 > 0.0.0.0.tftp: 23 RRQ "C009C815.SUN4C" octet

Wonder if anyone else has seen this ?. Can't find any into on how
tftp is initialised, under the hood, cryptic error message...

Chris

[YTC#1]

On 13/09/2020 19:34, Chris wrote:
> Strange fault with tftp on Solaris 10. Originally
> set up for an install server a couple of years
> ago, which did work. Added case for a Sun 4C
> machine, Solaris 2.6 install.
>
> The files and setup seem ok, and the tftp.udp6 svc
> starts up and comes online. As soon as the client is
> pointed at it, it transitions to maintenance mode.
>
> Here's some log info
>
> * inetadm says:
>
> enabled maintenance svc:/network/tftp/udp6:default
>
> * /var/adm/messages says:
>
> "inetd[311]: [ID 702911 daemon.error] The inetd_start
> method of instance svc:/network/tftp/udp6:default
> isn't a regular file".

What does the XML show ?
vccfg -s /network/tftp/udp6 listprop | grep start
general/restarter fmri
svc:/network/inetd:default
inetd_start method
inetd_start/group astring root
inetd_start/timeout_seconds count 0
inetd_start/type astring method
inetd_start/use_profile boolean false
inetd_start/user astring root
inetd_start/exec astring
"/usr/sbin/in.tftpd -s /etc/netboot"

(This is from S11, not got a S10 example at the moment)

>
> ?????
>
> tftp boot dir is at: /export/tftpboot and this
> is known working for the original clients.
>
> Startup line in /etc/inet/inetd.conf has been
> updated to reflect that.
>
> Ip address for client is at 192.9.200.21
>
> * /export/tftpboot line for host says:
>
> C009C815 -> inetboot.SUN.4-75-solaris.26
>
> points to:
>
> 173280 Jul 16  1997 inetboot.SUN.4-75-solaris.26
>
> * /etc/hosts entry:
>
> 192.9.200.21 ss2sol26
>
> * /etc/ethers entry:
>
> 08:00:20:c0:11:07   ss2sol26
>
> * tcpdump ether host 08:00:20:c0:11:07 (trimmed date & time)
>
> rarp who-is ss2sol26 tell ss2sol26
> rarp reply ss2sol26 at ss2sol26
> IP ss2sol26.3549 > darkstar.tftp:  23 RRQ "C009C815.SUN4C" octet
> IP ss2sol26.3549 > 0.0.0.0.tftp:  23 RRQ "C009C815.SUN4C" octet
> IP ss2sol26.3549 > 0.0.0.0.tftp:  23 RRQ "C009C815.SUN4C" octet
> IP ss2sol26.3549 > 0.0.0.0.tftp:  23 RRQ "C009C815.SUN4C" octet
>
> Wonder if anyone else has seen this ?. Can't find any into on how
> tftp is initialised, under the hood, cryptic error message...
>
> Chris
>

--
Bruce Porter
"The internet is a huge and diverse community but mainly friendly"
http://ytc1.blogspot.co.uk/
There *is* an alternative! http://www.openoffice.org/

[Chris]

Thanks. Perhaps another utility provides more info, but

inetadm -l /network/tftp/udp6 says:

SCOPE NAME=VALUE
name="tftp"
endpoint_type="dgram"
proto="udp6"
isrpc=FALSE
wait=TRUE
exec="/export/tftpboot"
user="root"
default bind_addr=""
default bind_fail_max=-1
default bind_fail_interval=-1
default max_con_rate=-1
default max_copies=-1
default con_rate_offline=-1
default failrate_cnt=40
default failrate_interval=60
default inherit_env=TRUE
default tcp_trace=FALSE
default tcp_wrappers=FALSE
default connection_backlog=10

/export/tfptpboot has the correct link pointing to the
boot file and /etc/bootparams looks ok, but it never gets
that far, not even downloading the initial netboot image.

Starting to think it may be the ss2 firmware tftp request
packet not the right format, but even if so, it should not
effectively crash the tftp service ?. Not very robust...

Chris

[Chris]

On 09/25/20 11:02, YTC#1 wrote:

> On 13/09/2020 19:34, Chris wrote:
>> Strange fault with tftp on Solaris 10. Originally
>> set up for an install server a couple of years
>> ago, which did work. Added case for a Sun 4C
>> machine, Solaris 2.6 install.
>>
>> The files and setup seem ok, and the tftp.udp6 svc
>> starts up and comes online. As soon as the client is
>> pointed at it, it transitions to maintenance mode.
>>
>> Here's some log info
>>

Had another session with this today and deleted and reinstalled the
tftp and tftpr packages and it now works.

Thanks for the reply anyway...

Chris