Always about 8 minutes behind the latest block about bsc-archive-snapshot

[allada]

If you need this for production, i suggest using a different file system. ZFS requires a lot of tuning to make it work optimally. The config used in the file is designed to fucus more on disk space compression than speed.

EBS volumes Have too high of Io latency. You'll want to use instances with ephemeral drives. I4i is a good class.

Also, EBS is ungodly expensive, especially for IO2, when ephemeral instances are significantly cheaper and faster.

[harveyff]

ok thanks for the suggestion,i will try it

[harveyff]

Hi, allada, thank you for your suggestion. After my comparison, using im4gn.4xlarge is indeed cheaper than m6a.4xlarge+EBS(io2).

At present, I am going to replace the instance and re-sync. Do I need to start all the processes from the beginning?

[allada]

There's multiple ways you can do this... I usually snapshot the end (the os) or just run the script until it starts downloading then kill it.

Then use nc or ssh with zfs snap , zfs send and zfs recv to copy the dataset over... But I'm a power user, so it might be easier for you to just start over.

[harveyff]

Thanks for your help, after I changed the instance to im4gn.4xlarge, it synced to the latest block smoothly, and it seems to be working normally.