Blockchain disk space...

+2 votes
I have been running a test environment with multichain the last month and I noticed that the blockchain is growing rather quickly. The size of the chain is well over 100MB in disk space in just a few weeks. I am wondering is this normal or is there something wrong with my setup?

In just a few weeks I have over 100,000 blocks in the chain. In the test environment I only have a dozen or so addresses with less than 100 transactions and about 4 nodes mining. Right now I am averaging around 200 megs per month that will add up to several Gigs in space per year if my calculations are correct. What can I expect in disk space after a year in operation with 100 addresses and well over 100 transactions per month?
asked Sep 19, 2016 by dtarsio

1 Answer

+1 vote

In your case, all of the space is being taken up by the empty blocks, rather than those few transactions you have initiated, which will take up less than a megabyte in total.

Each block has an 80-byte header, as well as a "coinbase" transaction that contains the miner's signature, even if there are no 'real' transactions in it. And those blocks are indexed in various ways.

So our recommendation would be to run a blockchain with a longer target-block-time parameter. We are also considering adding a future option whereby a block will not be mined if there are no real transactions within.

answered Sep 20, 2016 by MultiChain
I deleted the .log file and started the chain again.  In only a few hours its back up to several Gigs.  Its a huge file again but I was able to extract some data out of it.  Its logging about every half second "CanSend: FALSE". We are going to need some kind of fix as soon as possible because the .log file grows so large and so fast its maxing out disk space on most of the nodes causing them to shut down.
OK, thanks. Can you please grab the last 100000 lines of the file using "tail -n 100000 permissions.log" then gzip it and send to multichain.debug@gmail.com.
We got it - thanks - and are taking a look at the problem.
I'm told by our dev team that this has been fixed in alpha 26 already.
Ok I am running alpha 24.  I will upgrade and see if it fixes the problem.
...