Fixed commit to db every few blocks

[imperx]

While running syncnode.py I realised it is not committing fetched blocks to blockchain.db. Eventually it caused crashes if done on remote server.

[imperx]

lol no

I love the technical comment. If you start fetching db from block 0, eventually it crashes (memory issue? I don't know). The problem I encountered is easy to recreate. Just git clone the repo, and run syncnode. I did that on Amazon cloud and every time it crashed, needing to reboot the instance. Not saying, that locally, if it fetched 14 000 blocks, and crash, you have to start over from block 1. Look at Jack's earlier commit: 992c339 . I still think counter%1 == 3 makes no sense, unless you want the equation never be reached

[jacklevin74]

It is true there is a bug, but it’s not severe, the waiting to comment until 3 block records is in a buffer was mostly an experiment in batch commits. I’ll have to rewrite it to be more robust, for now people can experiment by varying this number to control batch size

…

On Tue, Oct 3, 2023 at 8:22 PM imperx ***@***.***> wrote: lol no I love the technical comment. If you start fetching db from block 0, eventually it crashes (memory issue? I don't know). The problem I encountered is easy to recreate. Just git clone the repo, and run syncnode. I did that on Amazon cloud and every time it crashed, needing to reboot the instance. Not saying, that locally, if it fetched 14 000 blocks, and crash, you have to start over from block 1. Look at Jack's earlier commit: 992c339 <992c339> . I still think counter%1 == 3 makes no sense, unless you want the equation never be reached — Reply to this email directly, view it on GitHub <#25 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAEEDZEAHP3KVLISX3HDSGTX5TI7JAVCNFSM6AAAAAA5RZQCFWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTONBWGA3TCNBTGU> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[nukee86]

Well, it crashed my EC2 instance few times. I realized the problem is only if you have to start over with downloading whole db - if there will be any interruption, you have to start over. It's not only me, someone in telegram group mentioned the problem aswell, I guided him with the solution and it worked. Hence I thought I'll mention it in case someone encounters the same problem.