Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Weaver]

One of my ZyXEL VMG1312-B10A modems running johnson custom firmware (written by our very own Mr johnson) has decided this morning to stop giving out FEC error count data points in the graph of errors vs time. It is drawing the graph grid lines and the values on the two axes, but there’s nothing showing on it, ie it’s either not drawing anything, or all FEC count values are zero. I logged in to the modem with telnet and entered the command ‘reboot’ which seemed to be successful in that it produced the expected prompts and then seemingly went down for a minute or so (maybe less). Was that the correct command to completely reboot the o/s ? I’m wondering if I perhaps just restarted the ZyXEL shell, no? I ask about this because the ‘reboot’ didn’t cure the problem.

The Mongoose web server is producing CRC count graphs as expected and all other graphs are as expected. It’s just FEC error counts. I notice that the logfile /data/logfile begins with two lines that just read "NOSYNC" alone, then after those two lines we have records as normal with the usual fields as expected.

Any idea how to fix this problem? Am I ok to just rename the logfile out of the way, in case its content is causing unhappiness? Maybe then perhaps delete the renamed file, to reclaim space.

I’m not able to power-cycle the device at the moment, and in any event I’m a bit wary about doing that. If it does turn out to be necessary to power-cycle the device, should I be entering any command to shut the system down cleanly before I then power-cycle it?

[burakkucat]

I logged in to the modem with telnet and entered the command ‘reboot’ which seemed to be successful in that it produced the expected prompts and then seemingly went down for a minute or so (maybe less). Was that the correct command to completely reboot the o/s ? I’m wondering if I perhaps just restarted the ZyXEL shell, no? I ask about this because the ‘reboot’ didn’t cure the problem.

That was a full software reboot of the device.

The Mongoose web server is producing CRC count graphs as expected and all other graphs are as expected. It’s just FEC error counts. I notice that the logfile /data/logfile begins with two lines that just read "NOSYNC" alone, then after those two lines we have records as normal with the usual fields as expected.

Any idea how to fix this problem? Am I ok to just rename the logfile out of the way, in case its content is causing unhappiness? Maybe then perhaps delete the renamed file, to reclaim space.

Login as the all-powerful "supervisor" user. At the Zyxel prompt issue a "sh" (or "shell") command to gain a BusyBox shell. Change directory to the data directory and edit the logfile to remove the two "NOSYNC" lines. Finally, issue a "reboot" command (from the BusyBox shell).

[Weaver]

Many thanks. I was exhausted today (Tue) and didn’t get a chance to execute your recipe before I found that the disease had spontaneously vanished. I’m assuming that the two lines in question caused the bug then, and they got pushed out in time as the file is a moving window representing only the last n seconds of recent time.

Any idea how it was caused in the first place?

I think a little awk program is in order to remove such a pair of lines.

[burakkucat]

Any idea how it was caused in the first place?

Sorry, no.

If you just wanted to truncate the file to size zero bytes, then

# > /data/logfile
# reboot

(from the BusyBox shell, as the supervisor) would do the deed.