Topic: Meteobridge server was down last night

[miraculon]

Here is a posting on the forum:

http://forum.meteohub.de/viewtopic.php?f=56&t=9569

Meteobridge Server Outage

Postby admin » Wed Jan 08, 2014 10:31 am
This is to inform you about the server outage that happened yesterday (server
is up and running in the meantime) and what I am doing to get this solved for
the future.

I host the Meteobridge server that provides the boot images for the Meteobridge
clients and some processing features (storing long term min/max values, web cam
image uploads and weather data annotations) at one of the well reputated hosting
companies here in Germany. It is Hetzner, where I also do colocation of my special
hardware Meteoplug server and also have generic root servers for web presence and forum
software. For Meteobridge I decided on a virtualized server as one can expect professional
desaster recovery as part of the virtualization concept (this takes some of the
burden from my shoulders) and no risks of longer outages by failed HW as
virtual servers can be moved in no time to new Hosts and continue running there.
Well this is the theory, but Hetzner demonstrated so far, that their VServers
are not ready for prime time. I am sorry for that.

The server is now up again, but I am just building a virtual server for Meteobridge
in North America. As this time it is a large provider with world wide reach, I am confident
not to have these outages any more and to get the benefits of virtualization not just on paper.

To make your Meteobridge independent in operation from the server you can add
a FAT32 formatted USB stick to it. Meteobridge will then store the needed images
on next boot on the stick and from then on it will boot from USB stick if server cannot be reached
after several tries. In that case some special features like payment function for Meteobridge purchase,
etc will not be functional but the normal operation will be fine without a server being online.

Sorry for the inconvenience, but at least I wanted to explain what happened and what I am
doing to solve this issue.

I got caught by this and panicked reverting to defaults and eventually reflashing the unit to no avail. I could have been a little smarter about it and tried pinging the server. Oh well. I got it all back running this morning.
I think that I will follow the advice about the USB stick, and also back up my settings this time

The symptom on my side was reverting to the default landing page no matter what I tried.

Greg

[saratogaWX]

I had the same issue, and sent an email to Boris -- he got the server back running promptly.

I also added a FAT32 formatted USB memory to the TP-LINK TL-MR3020 that runs Meteobridge .. as he says, this should prevent the issue in the future if the config server is not available when the device reboots.

What I don't know is whether the USB stick can be plugged into the hub used for the USB2->USB1 adapter that a USB datalogger needs.  My setup uses TCP/IP to VirtualVP, so I can't tell if the save of the working firmware would work on a USB stick plugged into a hub or not.

Best regards,
Ken

btw.. I have a new testing site on e-rice.net at www.saratogawx.net and it's driven by Meteohub

[docbee]

The stick can be plugged into the USB hub connected to the Meteobridge. That way the hub has something useful to do apart from translating USB 1.x to 2.0

To have the stick initialized Meteobridge will need one boot with the stick attached. Meteobridge will store the imaged received from the server to the stick and can then make use of it when server is not available.

BTW: I just finished migration to a new server in North America. It should be even more responsive for you than the previous one here in Germany and I expect the outages to be history as well.