simonacca / zatt Goto Github PK

Hi,I come with a problem when i RUN firstly.
arg --id is not exsited

https://github.com/simonacca/zatt/blob/master/README.md

"recording has been archived"

the master branch does not contain the --id argument:

zattd -c zatt.conf --id 0 -s zatt.0.persist --debug
zattd: error: unrecognized arguments: --id 0

however the feature/pickle seems to work. maybe you can upmerge to master?

Hello there! I'm trying to use Zatt for a project and noticed that when running
zattd -c zatt.conf --id 0 -s zatt.0.persist --debug
zattd throws an error that --id is an unrecognized argument. Turns out, "--id" is not even a flag recognized in server/config.py.

I'm currently looking into workarounds. One that didn't work was providing a node (from zatt.conf) IP addr and port in the --address and --port directories. I got the following error stacktrace for command zattd -c zatt.conf --address "127.0.0.1" --port 5254 -s zatt.persist --debug:

  File "/usr/local/bin/zattd", line 11, in <module>
    load_entry_point('Zatt==1.0', 'console_scripts', 'zattd')()
  File "/usr/local/lib/python3.6/site-packages/Zatt-1.0-py3.6.egg/zatt/server/main.py", line 31, in run
  File "/usr/local/lib/python3.6/site-packages/Zatt-1.0-py3.6.egg/zatt/server/main.py", line 10, in setup
  File "/usr/local/lib/python3.6/site-packages/Zatt-1.0-py3.6.egg/zatt/server/config.py", line 40, in __init__
  File "/usr/local/lib/python3.6/site-packages/Zatt-1.0-py3.6.egg/zatt/server/config.py", line 72, in _get
  File "/usr/local/lib/python3.6/site-packages/Zatt-1.0-py3.6.egg/zatt/server/config.py", line 72, in <setcomp>
ValueError: not enough values to unpack (expected 2, got 1) 

I'm going to look at the config.py file and see if there are any changes that could remedy this usage error. Wanted to give you a heads up in case you know anything about the problem.

Hi!

I've been reading some of zatt's code as I implement a toy in-memory key-value store with raft. Thank you for writing very readable code -- this has been a very helpful resource!

I notice that in zatt/server/states.py, you reset the election timer when the node is in a follower state before checking if the leader was up-to-date. This probably doesn't make much of a difference in practice, but wouldn't it be better to wait to reset the timer until we know that the leader was up-to-date? That way, if it isn't (i.e. the node has a more up-to-date term than the current leader) the node doesn't postpone starting an election.

It looks to me that this is how it is handled in Ongaro's pseudocode,