Hi all,
a while ago I started an additional Witness Server in order to double the numbers of snapshots created daily and therefore make the whole restore process a little faster.
Current situation is that @primersion has a daily snapshot running at 11 am which gets published and is available over his website under snapshots.
In case of an incident or a fork (the monster of fork seems quiet right now, thank god) all affected witnesses need to download a snapshot, restore and then catch up to the always advancing head block of the chain, the older the snapshot is and/or the slower the used hardware (nvme for the MongoDB rules) the longer it takes until the witness catches up.
But, the faster enough witnesses are back to verify blocks, the faster the growing backlog gets worked through and pending transactions get executed.
So, in short, I will use my backup witness server to create one additional snapshot each day, with 12 hours difference from the one @primersion is creating. Hosting will be done on one of my internet-facing servers, once the process runs automated and stable the snapshot will also be available over the site of @primersion
Server specs
That's not really a beast, but it's going to be enough thanks to the NVME where the MongoDB is located. At the core is a 6 Core Ryzen 1600, 16GB RAM, and a Gigabyte 1TB NVME running on PCIe3x (so there would be room to improve but the MB does not have x4 capabilities)
output of lswh -short
ETA
Well, it is like with every project in the IT world, it is done when it is done :)
My most educated guess would be something like Thursday or Friday this week. I will let you know once the process is stable.
Hive Engine Witness
As said above, I also run a normal witness for the Hive Engine Sidechain, currently ranked at #19
Happy about every vote, if you also want to leave a vote, @primersion 's site is also capable of doing that ;)
Thanks a lot for reading!
Have a nice evening,
Jan