telnetd: bind: address already in use, while not supposed to run at all?

[rjh2805]

In my ER logs I periodically see the following message popping up:

Aug  1 16:29:08 ubnt-er-poe telnetd[17518]: bind: Address already in use

 This surprises me as I haven't enabled the Telnet daemon, i.e., I don't have anything related to Telnet configured inside the 'service' tag of my config. However, when I check for the telnetd process, I see the following:

admin@ubnt-er-poe# ps fax | grep telnet
  584 ?        Ss     0:00 /usr/sbin/telnetd -p 55523 -b 127.0.0.101
15700 pts/0    S+     0:00                      \_ /bin/busybox grep telnet
[edit]
admin@ubnt-er-poe#

 So it seems telnetd is active. What does the error message in /var/log/messages actually point it? Any suggestions?

[UBNT-ancheng]

Yeah this has been discussed before (for example here) and the telnetd instance listening on 127.0.0.1 is for the implementation of the CLI window in the Web UI and the message is likely caused by reduntant starts of the daemon (which we should look into and fix).

[budcar]

Hi ancheng,

Newly up and running live, I'm monitoring my logs pretty closely.  I got the "telnetd bind" message also, and it just happend to follow some unsucessfull WAN attemtps at port 23: 

Aug 30 18:10:09 RT kernel: [WAN-LOCAL-default-D]IN=eth0 OUT= MAC=24:a4:3c:3d:xx:xx:xx:00:f1:40:01:d9:08:00 src=218.82.171.180 DST=(My IP Addr From ISP) LEN=60 TOS=0x00 PREC=0x00 TTL=52 ID=41647 DF PROTO=TCP SPT=41882 DPT=23 WINDOW=5440 RES=0x00 SYN URGP=0

Aug 30 18:10:12 RT kernel: [WAN-LOCAL-default-D]IN=eth0 OUT= MAC=24:a4:3c:3d:xx:xx:xx:00:f1:40:01:d9:08:00 src=218.82.171.180 DST=(My IP Addr From ISP) LEN=60 TOS=0x00 PREC=0x00 TTL=52 ID=41648 DF PROTO=TCP SPT=41882 DPT=23 WINDOW=5440 RES=0x00 SYN URGP=0

Aug 30 18:10:18 RT kernel: [WAN-LOCAL-default-D]IN=eth0 OUT= MAC=24:a4:3c:3d:xx:xx:xx:00:f1:40:01:d9:08:00 src=218.82.171.180 DST=(My IP Addr From ISP) LEN=60 TOS=0x00 PREC=0x00 TTL=52 ID=41649 DF PROTO=TCP SPT=41882 DPT=23 WINDOW=5440 RES=0x00 SYN URGP=0

Aug 30 18:10:41 RT telnetd[8896]: bind: Address already in use

I read your posts that the suspected cause is the support for the GUI CLI....in my case, at the time, i had no GUI connection to the ERpoe.

I was also wondering if you thought it is just coincidence that someone tried to hit port 23 within seconds of the error.  BTW, this is the first time in about 3 weeks of monitoring that i've seen this error in the log.

Thanks for any further insight you may have.

**EDIT: I have not configured the ER for telnet access

[UBNT-ancheng]

Yeah just coincidence, currently the telnetd instance is started at initialization time so does not require a GUI connection, but as discussed before we should probably only start it if GUI is in use. Again though note that this instance is listening on "localhost" so can only be accessed from the router itself. As mentioned we should look into the redundant starts of the daemon and fix that issue of course.

[budcar]

Ok thanks...i see now......i had totally spaced that the process was listening on 'localhost'.