3feca6d743
Since HF19, there is no longer a concept of a per node reward height, so we need a new sort criterion for active nodes. I believe the most interesting sort is that by most recent uptime proof. At the top of the list will appear nodes whose proof is overdue (i.e. > 1:00:00 ago). This should never be more than a few seconds, so one can see at a glance any nodes that are ripe for being moved to the Inactive/Decommissioned Service Nodes section the next time they are tested by the network. This also makes the list much more dynamic to watch. Since the Last Reward Height column is now redundant, we can replace it with something more interesting, such as the Active Since Height. |
||
|---|---|---|
| contrib | ||
| static | ||
| templates | ||
| .gitignore | ||
| base58.py | ||
| config.py | ||
| devnet.py | ||
| LICENSE | ||
| lmq.py | ||
| local_config.py | ||
| mainnet.py | ||
| observer.py | ||
| README.md | ||
| testnet.py | ||
Oxen Observer OMG block explorer
Block explorer using Oxen 8+ LMQ RPC interface that does everything through RPC requests. Sexy, awesome, safe.
Prerequisite packages
sudo apt install build-essential pkg-config libsodium-dev libzmq3-dev python3-dev python3-flask python3-babel python3-pygments python3-oxenmq
Note that the last requirement (python3-oxenmq) comes from the Oxen repository (https://deb.oxen.io).
Running in debug mode
To run it in debug mode (production requires setting up a WSGI server, see below):
FLASK_APP=observer flask run --reload --debugger
This mode seems to be a bit flakey, though -- reloading, in particular, seems to break things and make it just silently exit after a while, so only do this for quick and dirty testing.
Setting up for production with uwsgi-emperor:
Do the above, but instead of running it with flask directly, set up uwsgi-emperor as follows:
apt install uwsgi-emperor uwsgi-plugin-python3
in /etc/uwsgi-emperor/emperor.ini add configuration of:
# vassals directory
emperor = /etc/uwsgi-emperor/vassals
cap = setgid,setuid
emperor-tyrant = true
Create a "vassal" config for oxen-observer, /etc/uwsgi-emperor/vassals/oxen-observer.ini, containing:
[uwsgi]
chdir = /path/to/oxen-observer
socket = mainnet.wsgi
plugins = python3,logfile
processes = 4
manage-script-name = true
mount = /=mainnet:app
logger = file:logfile=/path/to/oxen-observer/mainnet.log
Set ownership of this user to whatever user you want it to run as, and set the group to _loki (so
that it can open the oxend unix socket):
chown MYUSERNAME:_loki /etc/uwsgi-emperor/vassals/loki-observer.ini
In the oxen-observer/mainnet.py, set:
config.oxend_rpc = 'ipc:///var/lib/loki/oxend.sock'
and finally, proxy requests from the webserver to the wsgi socket. For Apache I do this with:
# Allow access to static files (e.g. .css and .js):
<Directory /path/to/loki-observer/static>
Require all granted
</Directory>
DocumentRoot /home/jagerman/src/loki-observer/static
# Proxy everything else via the uwsgi socket:
ProxyPassMatch "^/[^/]*\.(?:css|js)(?:$|\?)" !
ProxyPass / unix:/path/to/loki-observer/mainnet.wsgi|uwsgi://uwsgi-mainnet-observer/
(you will probably need to a2enmod proxy_uwsgi to enable the Apache modules that make that work).
That should be it: restart apache2 and uwsgi-emperor and you should be good to go. If you want to
make uwsgi restart (for example because you are changing things) then it is sufficient to touch /etc/uwsgi-emperor/vassals/loki-observer.ini to trigger a reload (you do not have to restart the
apache2/uwsgi-emperor layers).
If you want to set up a testnet or devnet observer the procedure is essentially the same, but using testnet.py or devnet.py pointing to the oxend.sock from a testnet or devnet oxend.