You can't reboot via SNMP because SNMP doesn't support write currently. We have on our todo list a special URL you can query with a passsword to cause a reboot without needing the full web interface to load, but it's not going to be in a very near-term release. It will actually appear in the Base 3 code first, then get backported to the RackInjector code.
Your symptoms sound like the device is overloaded with broadcasts or other requests, is there something going on on that segment which would saturate the device? If you have a managed switch, make sure that the port that this is plugged into is relatively quiet for packets. One note: If you can get the page with the firmware reboot button to load enough to get the button then you can reboot. Often clicking on the tab which has the reboot button and then clicking on refresh will cause that part of the interface to load well enough to get that button up, since the web app remembers what page you were on when you click refresh. Just clicking on the tab will sometimes not work as the web app can get confused if the sitemonitor is behaving badly. The last bug we fixed had to do with the web server not freeing up resources. But that just caused dynamic variables to not load - the application would load, but wouldn't have any data. Doing the above procedure would allow remote reboots. I will take this moment to state one key recommendation: Set up the ping watchdog to point at a local address, probably the default gateway. After doing this, If you can't get into the interface to reboot the unit, shut the down the interface the rackinjector management interface is attached to (or otherwise prevent pings from getting through) and the unit will reboot automatically after the ping watchdog time. Note that this won't affect attached radios, except that on early boards (previous to revision B control boards), sync will drop for a couple seconds while the unit reboots. Also, we'd like to see the logs after any event like this, unless you aren't running recent firmware. We've got the code instrumented well enough that often the cause will show in the log. On Tue, Mar 31, 2020 at 11:21 PM George Skorup <gsko...@surfbroadband.com> wrote: > Got a RackInjector that's acting weird. I believe it's on the latest > version. But ICMP latency to it is all over the place (vs others that > respond 1ms or less continuously). HTTP responds-ish, sometimes. Pages > won't fully load. Then it will eventually not respond at all. Give it a > while and it will half respond again. But SNMP seems fine, at least > get/walk. I looked through the MIB and couldn't find anything that stood > out for reboot. Am I just not finding the correct OID, or is a reboot via > SNMP not possible? I would rather not like to send someone to power cycle > it, not to mention all of the gear powered by it. > -- > AF mailing list > AF@af.afmug.com > http://af.afmug.com/mailman/listinfo/af_af.afmug.com > -- - Forrest
-- AF mailing list AF@af.afmug.com http://af.afmug.com/mailman/listinfo/af_af.afmug.com
