Hi Grant,
No luck again with the w/a you provided.
Here is the result of more investigation:
1) disabled (disable = yes) in /etc/xinetd.d/tftp
2) reboot or just "service xinetd restart"
3) run in.tftpd
Writing to a new file still fails. This is what I discovered from / var/log/messages:
Sep 13 23:56:39 chl1 xinetd[2301]: START: tftp pid=3036 from=10.58.2.159
That is, xinetd automatically starts TFTP when a TFTP request comes in. 10.58.2.159 in fact is the IP address of the switch from which I'm trying to TFTP-ing.
Checking the processes it turns out that:
[root@chl1 log]# ps -ef | grep tftp root 3014 2813 0 23:55 pts/2 00:00:00 ./in.tftpd -c -v -u root -s /tftpboot root 3036 2301 0 23:56 ? 00:00:00 in.tftpd -s /tftpboot root 3043 2888 0 23:57 pts/3 00:00:00 grep tftp [root@chl1 log]#
two tftpd are running. Apparently the one responding is the one w/o the '-c' option!!!
Can someone please explain to me why xinetd is starting tftp, even if disabled!?!?!?
Thanks, Davide
On Sep 13, 2007, at 7:52 PM, Grant McChesney wrote:
On 9/13/07, Davide Grandis davide.grandis@fastwebnet.it wrote: Hi Grant,
Thanks for the support.
I had this same problem when trying to back up my switch configs. After some googling, I found a workaround for RH-based systems. The workaround is disable tftp in xinetd, and run in.tftpd manually. I never figured out why it would not work with xinetd.
May I ask you to tell me the detailed steps to achieve that... I'm an absolute beginner! :-)
Thanks again, Davide
Here's the post about this problem on fedora forum: http://forums.fedoraforum.org/archive/index.php/t-109735.html
Here's the ugly workaround:
In /etc/xinetd.d/tftp, set disable = yes Restart xinetd (/etc/init.d/xinetd restart) Run in.tftpd (for example, "/usr/sbin/in.tftpd -l -c -v -u root -s / tftpboot") You can add the previous command to /etc/rc.local to make it start on boot. _______________________________________________ CentOS mailing list CentOS@centos.org http://lists.centos.org/mailman/listinfo/centos