Upgrading to 1.35.3 (all 3 nodes down)

Staking & Delegation Operate a Stake Pool
1

image
image1365Ɨ815 95.2 KB

cardano-node 1.35.3 - linux-x86_64 - ghc-8.10
git rev 950c4e222086fed5ca53564e642434ce9307b0b9

cardano-cli 1.35.3 - linux-x86_64 - ghc-8.10
git rev 950c4e222086fed5ca53564e642434ce9307b0b9

Should I download the DB again?

1 Like
2

I have a 4th server that is not even loading up

image

1 Like
3

I would wait before upgrading a bit. There is no rush yet.

1 Like
4

Theyā€™re already upgraded and also weā€™re told to upgrade by IOHK. @Alexd1985 I deleted the DB on all 3 servers to try to get them to redownload.
image

1 Like
5

Itā€™s only for testing at the moment. I donā€™t think their messaging was very clear. Donā€™t rush, there is no panic to upgrade yet.

1 Like
6

If I recall correctly there may be a need to rescan the database when upgrading to 1.35, so perhaps that may take a few hours on mainnet in particular.

Also, if upgrading from 1.34.1 I believe that config files have not changed. However, the new config files from Environments - The Cardano Book seem to have different file names (the mainnet- prefix seems missing). Depending on how you handle the change, there are a few values that may need editing in the config.json file:

AlonzoGenesisFile "alonzo-genesis.json"
ByronGenesisFile "byron-genesis.json"
ShelleyGenesisFile "shelley-genesis.json"

I believe the only other considerations for upgrading to 1.35 are:

I hope that helps,

CHG

1 Like
7

No, you donā€™t need to download the DB again. But the db/immutale folder is being scanned again and the ledger database is being rebuilt again during this upgrade. This takes a few hours, depending on your hardware configuration. For the best possible hardware this takes 2-3 hours, for regular hardware, it can take more than 12 hours.
You should have upgraded one node, and when the upgrade was done, the upgrade on the other nodes should have been like this:

  • stop cardano-node on the upgraded server and on the next server to be upgraded

  • delete the contents of the db folder (everything except the immutable folder) from the server you want to upgrade

  • copy the contents of the db folder from the upgraded server to the server you want t upgrade (except the immutable folder, which you already have, and which takes most of the space), preferably using rsync.

  • replace the binaries on the server you want to upgrade

  • start the cardano-node on both servers.

  • copy the

1 Like
8

U shouldnā€™t ā€¦ first time it will take more time to startā€¦ for my relay it took ~2hoursā€¦ just check inside glive if CPU and RAM/MEM infos are changing/increasing

1 Like
9

I have upgraded three machines now to 1.35.3 and I donā€™t delete anything in the db folder. In each case it took several hours for the db to be rebuilt and then it just continued to run flawlessly. I left the config files unchanged.

1 Like
10

Thatā€™s the way it should work. Had the same experience.

11

Can anyone send me the DB

12

There seems to be a download service for it:

1 Like
13

Yes, this is the way it works. But if you want to speed it up, you can do as I described it. Upgrade one machine, then upgrade the rest as I described it.

1 Like
14

Hi, i have the same problem, I have tried to rebuild it but the synchronization does not advance:

troubleshooting_1_35_3_cardano

2 Likes
15

As said over and over, 1.35.3 will not work on the testnet, at the moment. You have to stay on 1.35.2, there.

5 Likes