Commit Graph

17 Commits (b8dd1b20564183c6269ea03e2996ba778252ba35)

Author SHA1 Message Date
Michael Zanetti b8dd1b2056 Don't create indices for the databases
Indices are really helpful on databases that hold many thousands
of rows and are searched frequently. On the other hand, indices
are harmful on tables which are frequently written to because not
only the actual entry needs to be updated but also the index.

So expecially in the particular case of the attributes tables,
the index is of more harm than use because we write updates all
the time and never read it. Removing the index on the attributes
table reduces the disk io footprint of the zigbee db by 30%.

As the other tables are mostly sitting still all the time, the
index doesn't really do much harm either in terms of disk io,
but it increases the database size, while not providing much
benefit, as those entries are also only read once at startup
and the tables are reasonably small (a few hundreds).
2022-10-08 00:44:49 +02:00
Simon Stürz f57ce01664 Update node initialization and finish the initialization in any case with the given information 2020-12-09 17:20:50 +01:00
Simon Stürz 2ed1487053 Handle network address changes and remove unrecognized nodes from the network once and for all if they show up again 2020-12-09 13:33:03 +01:00
Simon Stürz 3997b5a5de Make udev optional and provide basic cluster information in node object 2020-12-04 14:15:35 +01:00
Simon Stürz a651ece1cf Add multistate input cluster and improve internal endpoint initialization 2020-11-26 12:01:26 +01:00
Simon Stürz 5c3470e9d3 Update out of spec cluster handling and fix debian control 2020-11-16 12:11:35 +01:00
Simon Stürz 3d62162f5e Improve debug prints and rename adapter members 2020-11-01 22:17:59 +01:00
Simon Stürz db146a38d0 Remove settings and leave that to the stack using application 2020-10-29 11:45:28 +01:00
Simon Stürz 531d8075d4 Fix database loading error for node model identifier 2020-10-28 08:18:00 +01:00
Simon Stürz 5cc83bb956 Update database and add lqi and timestamp for nodes 2020-07-01 15:14:34 +02:00
Simon Stürz ed5517fbf1 Implement reply timeout and restructure uart communication queue 2020-06-30 16:49:09 +02:00
Simon Stürz cdd7a1a1aa Implement power configuration cluster 2020-06-13 21:18:01 +02:00
Simon Stürz 0992028a8a Rework entire uart communication flow and prepare level cluster command execution signals 2020-06-06 14:09:11 +02:00
Simon Stürz b3f5a13548 Move entirely to db storage and fix attribute loading 2020-06-05 13:11:33 +02:00
Simon Stürz 9dc50acbd5 Finalize first db schema without redundant information 2020-06-05 09:18:33 +02:00
Simon Stürz 5085487bda Implement first attempt for network database loading and saving 2020-06-04 21:58:21 +02:00
Simon Stürz 7cdf772b1a Start implementing clusters 2020-05-06 12:33:26 +02:00