Closed pablosymc closed 1 year ago
Snapshots are just made from running Kupo with a wildcard pattern. I have been a bit lazy to redo more recent snapshots for latest releases. Ideally, I would setup a nightly build for that to keep snapshots in sync.
Thanks @KtorZ Do you mind sharing the commands you used if possible?
kupo \
--node-config $CFG/$NETWORK/cardano-node/config.json \
--node-socket /tmp/node.socket \
--workdir $HOME/.cache/kupo/$NETWORK \
--since origin \
--match '*' \
--defer-db-indexes \
--prune-utxo
That's usually what I run for snapshots ☝️. With NETWORK
and CFG
being set to whatever they need to.
kupo \ --node-config $CFG/$NETWORK/cardano-node/config.json \ --node-socket /tmp/node.socket \ --workdir $HOME/.cache/kupo/$NETWORK \ --since origin \ --match '*' \ --defer-db-indexes \ --prune-utxo
That's usually what I run for snapshots ☝️. With
NETWORK
andCFG
being set to whatever they need to.
I was referring to the commands used to generate the snapshots :) Let's say I'm at the tip, and now want to create snapshots from where I'm at. What's the best way to go about that, so incase I need to do a quick restore I can?
Is it just a simple tar gz zip of the kupo.sqlite3
file (chunked into multiple files)?
Is it just a simple tar gz zip of the kupo.sqlite3 file (chunked into multiple files)?
Indeed! Here are the steps I usually go through (assuming a kupo.sqlite3
in .
)
sqlite3 ./kupo.sqlite3
> VACUUM;
> PRAGMA OPTIMIZE;
GZIP=-9 tar cvzf kupo.sqlite3-$NETWORK.tar.gz kupo.sqlite3
md5 kupo.sqlite3-$NETWORK.tar.gz
> 500 MB
, split the archive in chunks <= 500 MB
split -b 500M kupo.sqlite3-$NETWORK.tar.gz
And that's it. Could be nice to add those steps to the Makefile now that I think about it 🤔 (e.g. make snapshot
)
Exactly what I was looking for @KtorZ !! Thank you so much for your help and your time!
Hi, thanks so much for this awesome tool! This is kind of a selfish request, but I was wondering if we could get some more recent snapshots for (2.4.1)? Look's like the snapshots from 2.3.4 start at slot ~81m and it would just honestly just make syncing much quicker since we're at slot ~97m now?
Or if you have advice on how I can generate my own snapshots, I could also attempt that as well. Thanks so much for your time!