[bisq-network/projects] Migrate datastores to real database (#29)

chimp1984 notifications at github.com
Thu Aug 6 15:47:22 UTC 2020


Migrating to any DB solution will be a pretty large task. I recommend to take the more low hanging fruits first, specially as those might remove many of the main reasons for the need of a DB. 
E.g. We don't need to keep 15 MB of trade statistics (goes back to 4 years old data). Loading historical data only on demand would reduce that issue by 95%.

DB will not solve magically all problems. If you do not keep tyhe data in RAM you will move to a async model which can become very complex when it comes to the DAO handling of the DaoState, which will become a major issue as it is alreayd 75 MB. But solving that is a pretty difficult and high risk task. Maybe a DB will help but maybe it will require a asynchronous (multithreaded) setup which adds a lot of risks and complexity....

Just rough thoughts, I am not against it, just want to keep expectations in reasonable limits and point out possible problems.

Also what was discussed above regarding lost payout tx data: That is very likely not related at all, as pending trades data are small. It might be that we do not update/write the data at the moment when the payout tx is set, and if the user shut down quickly or has a crash the data might be lost. Is likely a one-liner to fix if one investigate (maybe I will do if I find time).

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/bisq-network/projects/issues/29#issuecomment-670010936
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.bisq.network/pipermail/bisq-github/attachments/20200806/48bcd7ed/attachment.html>


More information about the bisq-github mailing list