vitelabs/go-vite

Gvite is syncing, but very slowly, like 15blocks/sec. What can I do to improve?

Closed this issue · 3 comments

aaronllowe commented

I have a 16G swap file for mem, 1 core xeon, and 30gb of storage on aws.

roymoro commented

the bandwidth should be enough.

aaronllowe commented

But its sooooooooooooooooooooooooooooooooooooo slow!, also, after a while, (i dont know what happenes), on the Full Node stats page on explorer.vite.net, it says the block prop latency iis like 2 million ms. I end up having to restart my whole node, and its just not good. I have it running off of tmux, which makes it so you can run it while the ssh window is closed. Is there another way to do this?
Thanks

jerry-vite commented

But its sooooooooooooooooooooooooooooooooooooo slow!, also, after a while, (i dont know what happenes), on the Full Node stats page on explorer.vite.net, it says the block prop latency iis like 2 million ms. I end up having to restart my whole node, and its just not good. I have it running off of tmux, which makes it so you can run it while the ssh window is closed. Is there another way to do this?
Thanks

You can cd your log directory, usually is ~/.gvite/testdata/runlog on linux.
Then choose the date named log directory, input the command grep -i "elapse" vite.log on terminal.
The output on terminal will show download time for sync, like following:

t=2019-03-19T13:51:52+0800 lvl=info msg="download <file subgraph_921601_925200> from 150.109.46.50:8484 elapse 3.181576073s" module=net/fileClient
t=2019-03-19T13:52:05+0800 lvl=info msg="download <file subgraph_925201_928800> from 150.109.46.50:8484 elapse 5.773482438s" module=net/fileClient

Some old log files maybe compressed, you can uncompress them by command gzip -d *.log.gz first.

The gap between file download is the handle the file downloaded, eg. extract blocks, verify, and insert to the Chain.

Please present the log info.