Re: tftp problem
On Ma, 04 iun 19, 16:45:39, ghe wrote: > > > > > > I edited its config file (/etc/default/atftpd) to point it at /tftpboot > > and not to use inetd, but a restart of the server didn't seem to make > > any difference. Rebooting the computer did, though, and it showed up in > > ps with the options I'd written in the config. "restarting the server" usually means restarting the whole system, you probably meant "service" here? How were you restarting it (the service)? > > It might also be nice if systemd wouldn't ignore the echo commands in > > the init shell script -- some of them have been put there to try to > > figure why something's not working. Maybe you weren't looking into the right place? Kind regards, Andrei -- http://wiki.debian.org/FAQsFromDebianUser signature.asc Description: PGP signature
Re: tftp problem
On 6/4/19 2:03 PM, ghe wrote: Sorry Mr Doe. Pressed the wrong button and sent to you instead of the list... > On 6/4/19 11:32 AM, john doe wrote: > >> Do you mind sharing how you get it sorted out? > Not at all. This 'trivial' software wasn't at all trivial for me to set > up properly -- lots of confusion, wrong assumptions, and futzing. Much > more trouble than configuring Cisco's (thoroughly documented) IOS... > > So, and IIRC, > > When I installed, it didn't show up in ps, but it was there when I tried > to get it to save a file. I think it must have been in inetd. I was > expecting it to be a daemon by default. > > And it was writing to /srv/tftp (as per the inetd config that I didn't > know existed). That took a bit of thinking, since I'd always thought > TFTP used /tftpboot. And I'd never heard of /srv/tftp. Some surfing the > web found useful info about that. > > When I replaced its tftp dir in /srv with a link to /tftpboot, it > started writing to where I expected it to. > > I edited its config file (/etc/default/atftpd) to point it at /tftpboot > and not to use inetd, but a restart of the server didn't seem to make > any difference. Rebooting the computer did, though, and it showed up in > ps with the options I'd written in the config. > > The program (as a daemon), systemd, my router, my firewall, and I are > all happy now. > > It might be an idea for somebody to add a bit more info to the man page, > and/or the web site, to explain just what needs to happen to make this > go. It's pretty straightforward when you know its secrets and gotchas. > > It might also be nice if systemd wouldn't ignore the echo commands in > the init shell script -- some of them have been put there to try to > figure why something's not working. -- Glenn English
Re: tftp problem
On 6/4/2019 6:27 PM, ghe wrote: > On 6/3/19 11:39 PM, john doe wrote: > > >> 'Bind' means use this adress only. > > Thanks. It means different things in different contexts. Wasn't sure > about this one. > >> 's.*d'? > > That was intended to be a joke -- a regex for systemd. I thought this > list would recognize it. Sorry. (atftpd's working now, BTW.) > Always interesting to learn something new!!! :) Do you mind sharing how you get it sorted out? -- John Doe
Re: tftp problem
On 6/3/19 11:39 PM, john doe wrote: > 'Bind' means use this adress only. Thanks. It means different things in different contexts. Wasn't sure about this one. > 's.*d'? That was intended to be a joke -- a regex for systemd. I thought this list would recognize it. Sorry. (atftpd's working now, BTW.) -- Glenn English
Re: tftp problem
On 6/3/2019 8:39 PM, ghe wrote: > On 6/3/19 11:48 AM, deloptes wrote: > >>> Response, run as root: >>> atftpd: can't bind port :69/udp >>> >> >> why do you have /udp after the port? Check the config > > 1) I don't; atftpd put it there. TFTP is UDP on port 69, isn't it? > Because in the command I gave in this thread '--port' option was missing. $ atftpd --port 2345 --verbose=7 --daemon --no-fork --logfile /dev/stdout [DIRECTORY] Port 69 UDP on the server and is also the destination port on the clients. > 2) What config? The closest thing to a config I could find was > /etc/default/atftpd. And it looked to me like it was ignoring that. > As far as I can tell, there is no config file per say but the init.d script will use the file '/etc/default/atftp'. >>> They're just kidding. I think -- I don't know what they mean by 'bind'. 'Bind' means use this adress only. > When I ran the file as a shell script, it chatted a bit. And most most > importantly, the daemon started and it started writing into the correct > directory. I need to see if I can get the s.*d word to do that. > 's.*d'? -- John Doe
Re: tftp problem
On 6/3/19 11:48 AM, deloptes wrote: >> Response, run as root: >> atftpd: can't bind port :69/udp >> > > why do you have /udp after the port? Check the config 1) I don't; atftpd put it there. TFTP is UDP on port 69, isn't it? 2) What config? The closest thing to a config I could find was /etc/default/atftpd. And it looked to me like it was ignoring that. >> They're just kidding. I think -- I don't know what they mean by 'bind'. >> I've saved configs over tftp from my Juniper firewall and my Cisco router. > > This usually means that the process can start listening on that port But if it means it can start listening, why say anything about it? They didn't tell me the sky is blue. And why not say any more? I was hoping for some useful info about what was wrong. I went into the startup script in init.d, hung a '.sh' on the filename, added some debugging echo statements, and commented out what I suspected was be making the script not print any of the echos. When I ran the file as a shell script, it chatted a bit. And most most importantly, the daemon started and it started writing into the correct directory. I need to see if I can get the s.*d word to do that. Thanks folks. You've given me info on a direction to go with this. -- Glenn English
Re: tftp problem
ghe wrote: > Response, run as root: > atftpd: can't bind port :69/udp > why do you have /udp after the port? Check the config > They're just kidding. I think -- I don't know what they mean by 'bind'. > I've saved configs over tftp from my Juniper firewall and my Cisco router. This usually means that the process can start listening on that port
Re: tftp problem
On 6/3/19 8:08 AM, john doe wrote: > atftpd --verbose=7 --daemon --no-fork --logfile /dev/stdout Response, run as root: atftpd: can't bind port :69/udp They're just kidding. I think -- I don't know what they mean by 'bind'. I've saved configs over tftp from my Juniper firewall and my Cisco router. Juniper a few minutes ago (small file; just beginning to get it going): > Save configurations (5213 bytes) to SSG140Config.txt on TFTP server 192.168.2.3 from ethernet0/7. > > tftp transferred records = 11 > tftp success! > > TFTP Succeeded ghe@sbox:~$ ls -lh /tftpboot/ | egrep SSG -rw-rw-rw- 1 nobody nogroup 5.1K Jun 3 08:48 SSG140Config.txt Cisco (bigger file; my border router): > > 12386 bytes copied in 0.584 secs (21209 bytes/sec) (Cisco makes '.'s instead of '!'s on failure.) ghe@sbox:~$ ls -lh /tftpboot/ | egrep Run -rw-rw-rw- 1 nobody nogroup 13K Jun 3 08:51 brouterRunningConfig.txt Those two have been known to be picky. And, as far as I know, there's no way to get them to use anything but port 69. I've also tftp'ed files locally, using atftp. ... > Also, is your " [directory] must be a world readable/writable directories."? ghe@sbox:~$ ls -lh / | egrep tftp drwxrwxrwx 2 nobody root4.0K Jun 2 16:00 tftpboot ... > drwxr-xr-x 2 nobody root 4096 /srv/tftp Right now, /srv/tftp is a link to /tftpboot, so it's 777: ghe@sbox:~$ ls -lh /srv | egrep tftp lrwxrwxrwx 1 root root 10 May 31 13:04 tftp -> /tftpboot/ It's owned by root, but it was nobody's when it was a dir (worked just fine then, anyway). And there seems to be no problem with the link being owned by root. -- Glenn English
Re: tftp problem
On 6/3/2019 4:08 PM, john doe wrote: > On 6/3/2019 3:24 PM, ghe wrote: >> On 6/2/19 11:46 PM, john doe wrote: >> >>> I assume that you have restarted the service? >> >> Yup. Several times on multiple computers. No joy. >> >> I haven't, though, tried starting with the options added by hand... >> > > $ atftpd --verbose=7 --daemon --no-fork --logfile /dev/stdout > > Should hopefully help you figure out what's going on. > Also, is your " [directory] must be a world readable/writable directories."? Just after installing 'atftpd": drwxr-xr-x 2 nobody root 4096 /srv/tftp -- John Doe
Re: tftp problem
On 6/3/2019 3:24 PM, ghe wrote: > On 6/2/19 11:46 PM, john doe wrote: > >> I assume that you have restarted the service? > > Yup. Several times on multiple computers. No joy. > > I haven't, though, tried starting with the options added by hand... > $ atftpd --verbose=7 --daemon --no-fork --logfile /dev/stdout Should hopefully help you figure out what's going on. -- John Doe
Re: tftp problem
On 6/2/19 11:46 PM, john doe wrote: > I assume that you have restarted the service? Yup. Several times on multiple computers. No joy. I haven't, though, tried starting with the options added by hand... -- Glenn English
Re: tftp problem
On 6/2/2019 9:47 PM, ghe wrote: > buster, atftp, atftpd > > atftpd writes to /srv/tftp, even when I change to /tftpboot in the > /etc/default config. Or, as is suggested in the man page, just leave > that part of the config empty. Restart doesn't impress it. > > Nice piece of software, but I have to put a link in /srv to /tftpboot to > make it do what tftp is supposed to do. Anybody seen this? Or know how > to fix it? > I assume that you have restarted the service? -- John Doe
tftp problem
buster, atftp, atftpd atftpd writes to /srv/tftp, even when I change to /tftpboot in the /etc/default config. Or, as is suggested in the man page, just leave that part of the config empty. Restart doesn't impress it. Nice piece of software, but I have to put a link in /srv to /tftpboot to make it do what tftp is supposed to do. Anybody seen this? Or know how to fix it? Works fine, I just don't like it when I ask it to do something and I have to create a little untidiness to make it happen... Buglet? PEBCAK? -- Glenn English
tftp problem with potato
Hi, after doing a new installation of debian potato on our server, i have a strange tftp problem: If i issue a tftp get command on a client, tftp hangs. If i do 'strace tftp', it works. Same if i strace in.tftpd on the server. Is this a weird timing problem? I use kernel 2.2.10, potato as of 12-JUL-1999. Other network functions seem to be ok. Thanks for any hints Heinrich Rebehn University of Bremen Physics / Electrical and Electronics Engineering - Department of Telecommunications - E-mail: mailto:[EMAIL PROTECTED] Phone : +49/421/218-4664 Fax :-3341 - The white sun glimmers on the mountain, the Huangho flows out to sea. To follow it over those thousand miles, I must rise one level more. --Wang Chi Huan
