When I click on this node from another in Mesh Status, I get this error. I am Running 3.19.3.0 on all our nodes.
I am encountering a problem with my Rockets running out of memory. This is from one node.
I've tried to reboot unit by ssh, but get:
This particular unit is an hour by Razor up a mountain, so I'd like to bounce it remotely, but may have to wait for a trip up this fall. Thoughts?
The others I can more easily get to to cycle the power. I'll throw a weekly reboot in crontab. But I hate to treat them like Windose machines<grin>.
With the current firmware I was having probs with nodes only showing IP address, so I took a suggestion to add a line to /etc/init.d/local:
/etc/init.d/olsrd restart
I don't know if this caused a problem.
73 Charlie KØTAN
Lake Havasu City, AZ
Mesh phone: 928 3001
I have 2 nodes now this morning that are displaying that error. I can not ssh into 1 of them at this time to reboot it, but I will do that when I get home.
We thought at one time the units (1 is a LITEBEAM, the other is a PowerBeam 5g and previously a Rocket M5) were locking up/running out of memory because we were not stopping the mesh status or wireless signal status refresh processes. We've made it a procedure to stop any active refresh before leaving the node, but it appears to not have solved that problem. These nodes are all running 3.19.3.0 except for perhaps the LITEBEAM which might be running a nightly build.
We've also found that if one can catch the issue early enough, the node will still respond to network activity via ssh. Today, its refusing my attempts to let me in.
I like the idea of a reboot via cron on a set period of time. Will check that out when I get closer to a node that is working.
Nick - VA7NIC
I have a 5.8 rocket (3.19.3.0) which is connected to a VLAN switch for a dTd brirdge out from our normal 2.4 mesh to a couple of new 5.8 devices. It has approx 60 days of uptime, but because it wasn't due for any expansion or work until the EOY, no one noticed that it had dropped out of sight. It was working correctly a month ago when I left town.
Now, it does not show up as a dTd link at all and a quick survey with a CPE510 in the near physical al area does not find it anywhere on RF. It does respond on a LAN http link but only gives the dredded "Internal Server Error Failed to create CGI process: Out of memory". I only had a short time to troubleshoot it, and I couldn't get in with Telnet/SSH (but might have been fat fingers) so I left it as is, rather than power cycling it.
It was not active for mesh activity and I don't think anyone left a status page process running (very limited access). While it does have a MeshChat instance that was previously running just fine across the dTd link, I know that that MeshChat zone was not actively updated in the past month.
So, it appears to have very quietly run out of memory and I'd like to know why.
Could someone please give me some ideas on how to query the true status of this device (from LAN) and hopefully grab some good data on current state. It obviously can't create a CGI process but the resident webserver is still serving up the local error message, so something must still be alive.
Any/all advice appreciated. At some point I will need to power cycle it and get it back on the air, but it would be good to retrieve some data first.
TIA,
- Don - AA7AU