Upgrade Log
For each upgrade of the Oasis Network, we are tracking important changes for node operators' deployments.
They are enumerated and explained in this document.
2023-11-29 (10:02 UTC) - Eden Upgrade
For a detailed view of the Eden Upgrade process, please refer to the Eden Upgrade page.
The Oasis Core 23.0.x binary in our published releases is built only for Ubuntu 22.04 (GLIBC>=2.32). You'll have to build it yourself if you're using prior Ubuntu versions (or other distributions using older system libraries).
Configuration Changes
The node configuration has been refactored so that everything is now configured via a YAML configuration file and configuring via command-line options is no longer supported.
Some configuration options have changed and so the configuration file needs to be updated. To make this step easier, a command-line tool has been provided that will perform most of the changes automatically. You can run it with:
oasis-node config migrate --in config.yml --out new-config.yml
The migration subcommand logs the various changes it makes and warns you if a config option is no longer supported, etc. At the end, any unknown sections of the input config file are printed to the terminal to give you a chance to review them and make manual changes if required.
Note that the migration subcommand does not preserve comments and order of sections from the input YAML config file. You should always carefully read the output of this command, as well as compare the generated config file with the original before using it.
After you are satisfied with the new configuration file, replace the old file with the new one as follows:
mv new-config.yml config.yml
The configuration format for seed nodes has changed and it now requires the node's P2P public key to be used. In case your old configuration file contains known Mainnet seed nodes, this transformation is performed automatically.
However, if it contains unknown seed nodes then the conversion did not happen automatically and you may need to obtain the seed node's P2P public key. For Mainnet you can use the following addresses:
TBD
TBD
Please be aware that every seed node should be configured to listen on two distinct ports. One is dedicated to peer discovery within the CometBFT P2P network, while the other is used to bootstrap the Oasis P2P network.
Data Directory Changes
The subdirectory (located inside the node's data directory) used to store
consensus-related data, previously called tendermint
(after the consensus
layer protocol backend) has been renamed to consensus
in Oasis Core 23.0.x. If
any of your scripts rely on specific directory names, please make sure to update
them to reflect the changed sdirectory name.
State Changes
The following parts of the genesis document will be updated:
For a more detailed explanation of the parameters below, see the Genesis Document docs.
All state changes will be done automatically with the migration command provided
by the new version of oasis-node
. It can be used as follows to derive the same
genesis file from an existing state dump at the correct height (assuming there
is a genesis.json
present in the current working directory):
oasis-node genesis migrate --genesis.new_chain_id oasis-4
General
-
chain_id
will be set tooasis-4
. -
halt_epoch
will be removed as it is no longer used.
Registry
-
registry.runtimes[].txn_scheduler.propose_batch_timeout
specifies how long to wait before accepting proposal from the next backup scheduler. It will be set to5000000000
(5 seconds). Previously the value was represented in the number of consensus layer blocks. -
registry.params.gas_costs.prove_freshness
specifies the cost of the freshness proof transaction. It will be set to1000
. -
registry.params.gas_costs.update_keymanager
specifies the cost of the keymanager policy update transaction. It will be removed as the parameter has been moved underkeymanager.params.gas_costs.update_policy
. -
registry.params.tee_features
specify various TEE features supported by the consensus layer registry service. These will be set to the following values to activate the new features:"tee_features": {
"sgx": {
"pcs": true,
"signed_attestations": true,
"max_attestation_age": 1200
},
"freshness_proofs": true
} -
registry.params.max_runtime_deployments
specifies the maximum number of runtime deployments that can be specified in the runtime descriptor. It will be set to5
.
Root Hash
roothash.params.max_past_roots_stored
specifies the maximum number of past runtime state roots that are stored in consensus state for each runtime. It will be set to1200
.
Staking
-
staking.params.commission_schedule_rules.min_commission_rate
specifies the minimum commission rate. It will be set to0
to maintain the existing behavior. -
staking.params.thresholds.node-observer
specifies the stake threshold for registering an observer node. It will be set to100000000000
base units (or100
tokens), same as for existing compute nodes.
Key Manager
-
keymanager.params.gas_costs
specify the cost of key manager transactions. These will be set to the following values:"gas_costs": {
"publish_ephemeral_secret": 1000,
"publish_master_secret": 1000,
"update_policy": 1000
}
Random Beacon
beacon.base
is the network's starting epoch. It will be set to the epoch of Mainnet's state dump + 1,28017
.
Governance
governance.params.enable_change_parameters_proposal
specifies whether parameter change governance proposals are allowed. It will be set totrue
.
Consensus
consensus.params.max_block_size
specifies the maximum block size in the consensus layer. It will be set to1048576
(1 MiB).
Other
-
extra_data
will be set back to the value in the Mainnet genesis file to include the Oasis Network's genesis quote:”Quis custodiet ipsos custodes?” [submitted by Oasis Community Member Daniyar Borangaziyev]:
"extra_data": {
"quote": "UXVpcyBjdXN0b2RpZXQgaXBzb3MgY3VzdG9kZXM/IFtzdWJtaXR0ZWQgYnkgT2FzaXMgQ29tbXVuaXR5IE1lbWJlciBEYW5peWFyIEJvcmFuZ2F6aXlldl0="
}
2022-04-11 (8:30 UTC) - Damask Upgrade
- Upgrade height: upgrade is scheduled to happen at epoch 13402.
We expect the Mainnet network to reach this epoch at around 2022-04-11 8:30 UTC.
Instructions - Voting
Voting for the upgrade proposal will end at epoch 13152. We expect the Mainnet network to reach this epoch at around 2022-03-31 21:00 UTC.
At this time only entities which have active validator nodes scheduled in the validator set are eligible to vote for governance proposals.
At least 75% of the total voting power of the validator set needs to cast a vote on the upgrade proposal for the result to be valid.
At least 90% of the vote needs to be yes votes for a proposal to be accepted.
The Oasis Protocol Foundation has submitted an upgrade governance proposal with the following contents:
{
"v": 1,
"handler": "mainnet-upgrade-2022-04-11",
"target": {
"consensus_protocol": {
"major": 5
},
"runtime_host_protocol": {
"major": 5
},
"runtime_committee_protocol": {
"major": 4
}
},
"epoch": 13402
}
To view the proposal yourself, you can run the following command on your online Oasis Node:
oasis-node governance list_proposals -a $ADDR
where $ADDR
represents the path to the internal Oasis Node UNIX socket
prefixed with unix:
(e.g.unix:/serverdir/node/internal.sock
).
The output should look like:
[
{
"id": 2,
"submitter": "oasis1qpydpeyjrneq20kh2jz2809lew6d9p64yymutlee",
"state": "active",
"deposit": "10000000000000",
"content": {
"upgrade": {
"v": 1,
"handler": "mainnet-upgrade-2022-04-11",
"target": {
"consensus_protocol": {
"major": 5
},
"runtime_host_protocol": {
"major": 5
},
"runtime_committee_protocol": {
"major": 4
}
},
"epoch": 13402
}
},
"created_at": 12984,
"closes_at": 13152
}
]
Obtain your entity's nonce and store it in the NONCE
variable. You can do
that by running:
ENTITY_DIR=<PATH-TO-YOUR-ENTITY>
ADDRESS=$(oasis-node stake pubkey2address --public_key \
$(cat $ENTITY_DIR/entity.json | jq .id -r))
NONCE=$(oasis-node stake account nonce --stake.account.address $ADDRESS -a $ADDR)
where <PATH-TO-YOUR-ENTITY>
is the path to your entity's descriptor, e.g.
/serverdir/node/entity/
.
To vote for the proposal, use the following command to generate a suitable transaction:
oasis-node governance gen_cast_vote \
"${TX_FLAGS[@]}" \
--vote.proposal.id 2 \
--vote yes \
--transaction.file tx_cast_vote.json \
--transaction.nonce $NONCE \
--transaction.fee.gas 2000 \
--transaction.fee.amount 2000
where TX_FLAGS
refer to previously set base and signer flags as described in
the Oasis node CLI Tools Setup doc.
If you use a Ledger-signer backed entity, you will need to install version 2.3.2 of the Oasis App as described in Installing Oasis App on Your Ledger Wallet.
Note that the previous version of the Oasis App available through Ledger Live,
version 1.8.2, doesn't support signing the governance.CastVote
transaction
type.
To submit the generated transaction, copy tx_cast_vote.json
to the online
Oasis node and submit it from there:
oasis-node consensus submit_tx \
-a $ADDR \
--transaction.file tx_cast_vote.json
Instructions - Before upgrade
This upgrade will upgrade Oasis Core to the 22.1.x release series which
no longer allow running Oasis Node (i.e. the oasis-node
binary)
as root (effective user ID of 0).
Running network accessible services as the root user is extremely bad for
system security as a general rule. While it would be "ok" if we could drop
privileges, syscall.AllThreadsSyscall
does not work if the binary uses cgo
at all.
Nothing in Oasis Node will ever require elevated privileges.
Attempting to run the oasis-node
process as the root user will now terminate
immediately on startup.
While there may be specific circumstances where it is safe to run network services with the effective user ID set to 0, the overwhelming majority of cases where this is done is a misconfiguration.
Please, follow our Changing Your Setup to Run Oasis Services with Non-root System User guide for steps-by-step instructions on how to update your system.
If the previous behavior is required, the binary must be run in unsafe/debug
mode (via the intentionally undocumented flag), and debug.allow_root
must also
be set.
Instructions - Upgrade day
Following steps should be performed on 2022-04-11 only after the network has reached the upgrade epoch and has halted:
- Download the genesis file published in the Damask upgrade release.
Mainnet state at epoch 13402 will be exported and migrated to a 22.1.x compatible genesis file.
Upgrade genesis file will be published on the above link soon after reaching the upgrade epoch.
-
Verify the provided Damask upgrade genesis file by comparing it to network state dump.
The state changes are described in the Damask Upgrade document.
See instructions in the Handling Network Upgrades guide.
-
Replace the old genesis file with the new Damask upgrade genesis file.
-
Ensure your node will remain stopped by disabling auto-starting via your process manager (e.g., systemd or Supervisor)
-
Replace the old version of Oasis Node with version 22.1.3.
State of ParaTimes/runtimes is not affected by this upgrade and should NOT be wiped.
We recommend backing up the pre-Damask consensus state so all the transactions and history are not permanently lost.
Also, if you ever need to access that state in the future, you will be able to spin up an Oasis Node in archive mode and query the pre-Damask state.
-
Perform any needed configuration changes described below.
-
(only for ParaTime operators) Replace old versions of your ParaTime binaries with the new Oasis Runtime Containers (
.orc
) introduced in the Damask upgrade.For the official Oasis ParaTimes, use the following versions:
-
(only Emerald Web3 Gateway operators) Replace old version of Emerald Web3 Gateway with version 2.1.0.
-
(only Rosetta Gateway operators) Replace old version of Oasis Rosetta Gateway with version 2.2.0.
-
Start your node and re-enable auto-starting via your process manager.
Configuration Changes
To see the full extent of the changes examine the Change Log: of the Oasis Core 22.1.3, in particular the 22.1 and 22.0 sections.
If your node is currently configured to run a ParaTime, you need to perform some additional steps.
The way ParaTime binaries are distributed has changed so that all required
artifacts are contained in a single archive called the Oasis Runtime Container
and have the .orc
extension.
Links to updated ParaTime binaries will be published on the Network Parameters page for their respective ParaTimes.
The configuration is simplified as the runtime.paths
now only needs to list
all of the supported .orc
files (see below for an example).
Instead of separately configuring various roles for a node, there is now a
single configuration flag called runtime.mode
which enables the correct roles
as needed. It should be set to one of the following values:
none
(runtime support is disabled, only consensus layer is enabled)compute
(node is participating as a runtime compute node for all the configured runtimes)keymanager
(node is participating as a keymanager node)client
(node is a stateful runtime client)client-stateless
(node is a stateless runtime client and connects to remote nodes for any state queries)
Nodes that have so far been participating as compute nodes should set the mode
to compute
and nodes that have been participating as clients for querying
and transaction submission should set it to client
.
The following configuration flags have been removed:
runtime.supported
(existingruntime.paths
is used instead)worker.p2p.enabled
(now automatically set based on runtime mode)worker.compute.enabled
(now set based on runtime mode)worker.keymanager.enabled
(now set based on runtime mode)worker.storage.enabled
(no longer needed)
Also the worker.client
option is no longer needed unless you are providing
consensus layer RPC services.
For example, if your previous configuration looked like:
runtime:
supported:
- "000000000000000000000000000000000000000000000000000000000000beef"
paths:
"000000000000000000000000000000000000000000000000000000000000beef": /path/to/runtime
worker:
# ... other settings omitted ...
storage:
enabled: true
compute:
enabled: true
client:
port: 12345
addresses:
- "xx.yy.zz.vv:12345"
p2p:
enabled: true
port: 12346
addresses:
- "xx.yy.zz.vv:12346"
The new configuration should look like:
runtime:
mode: compute
paths:
- /path/to/runtime.orc
worker:
# ... other settings omitted ...
p2p:
port: 12346
addresses:
- "xx.yy.zz.vv:12346"
Additional notes
Examine the Change Log of the 22.1.3 release, in particular the 22.1 and 22.0 sections.
2021-08-31 (16:00 UTC) - Parameter Update
- Upgrade height: upgrade is scheduled to happen at epoch 8049.
We expect the Mainnet network to reach this epoch at around 2021-08-31 16:00 UTC.
Proposed Parameter Changes
The Oasis Core 21.2.8 release contains the consensus-params-update-2021-08
upgrade handler which will update the following parameters in the consensus layer:
- **
staking.params.max_allowances
** specifies the maximum number of allowances on account can store. It will be set to16
(default value is0
) to enable support for beneficiary allowances which are required to transfer tokens into a ParaTime. - **
staking.params.gas_costs
** ,governance.params.gas_costs
androothash.params.gas_costs
specify gas costs for various types of staking, governance and roothash transactions. Gas costs for transactions that were missing gas costs will be added. scheduler.params.max_validators
is the maximum size of the consensus committee (i.e. the validator set). It will be increased to110
(it was set to100
previously).
Instructions - Voting
Voting for the upgrade proposal will end at epoch 7876. We expect the Mainnet network to reach this epoch at around 2021-08-24 12:00 UTC.
At this time only entities which have active validator nodes scheduled in the validator set are eligible to vote for governance proposals.
At least 75% of the voting power needs to cast vote on the upgrade proposal for the result to be valid.
At least 90% of the votes need to be yes votes for a proposal to be accepted.
This upgrade will be the first upgrade that will use the new on-chain governance service introduced in the Cobalt Upgrade.
The Oasis Protocol Foundation has submitted an upgrade governance proposal with the following contents:
{
"v": 1,
"handler": "consensus-params-update-2021-08",
"target": {
"consensus_protocol": {
"major": 4
},
"runtime_host_protocol": {
"major": 3
},
"runtime_committee_protocol": {
"major": 2
}
},
"epoch": 8049
}
To view the proposal yourself, you can run the following command on your online Oasis Node:
oasis-node governance list_proposals -a $ADDR | jq
where $ADDR
represents the path to the internal Oasis Node UNIX socket prefixed with unix:
(e.g.unix:/serverdir/node/internal.sock
).
The output should look like:
[
{
"id": 1,
"submitter": "oasis1qpydpeyjrneq20kh2jz2809lew6d9p64yymutlee",
"state": "active",
"deposit": "10000000000000",
"content": {
"upgrade": {
"v": 1,
"handler": "consensus-params-update-2021-08",
"target": {
"consensus_protocol": {
"major": 4
},
"runtime_host_protocol": {
"major": 3
},
"runtime_committee_protocol": {
"major": 2
}
},
"epoch": 8049
}
},
"created_at": 7708,
"closes_at": 7876
}
]
Obtain your entity's nonce and store it in the NONCE
variable. You can do that by running:
ENTITY_DIR=<PATH-TO-YOUR-ENTITY>
ADDRESS=$(oasis-node stake pubkey2address --public_key \
$(cat $ENTITY_DIR/entity.json | jq .id -r))
NONCE=$(oasis-node stake account nonce --stake.account.address $ADDRESS -a $ADDR)
where <PATH-TO-YOUR-ENTITY>
is the path to your entity's descriptor, e.g. /serverdir/node/entity/
.
To vote for the proposal, use the following command to generate a suitable transaction:
oasis-node governance gen_cast_vote \
"${TX_FLAGS[@]}" \
--vote.proposal.id 1 \
--vote yes \
--transaction.file tx_cast_vote.json \
--transaction.nonce $NONCE \
--transaction.fee.gas 2000 \
--transaction.fee.amount 2000
where TX_FLAGS
refer to previously set base and signer flags as described in the Oasis node CLI Tools Setup doc.
If you use a Ledger-signer backed entity, you will need to install version 2.3.1 of the Oasis App as described in Installing Oasis App 2.3.1 to Your Ledger. This is needed because the current version of the Oasis App available through Ledger Live, version 1.8.2, doesn't support signing the governance.CastVote
transaction type.
To submit the generated transaction, copy tx_cast_vote.json
to the online Oasis node and submit it from there:
oasis-node consensus submit_tx \
-a $ADDR \
--transaction.file tx_cast_vote.json
Instructions - Before Upgrade System Preparation
- This upgrade will upgrade Oasis Core to version 21.2.8 which:
- Upgrades the BadgerDB database backend from v2 to v3. See BadgerDB v2 to v3 Migration section for required steps to be done before upgrade.
- Has a check that makes sure the file descriptor limit is set to an appropriately high value (at least 50000). While previous versions only warned in case the limit was set too low, this version will refuse to start. Follow the File Descriptor Limit documentation page for details on how to increase the limit on your system.
- Stop your node, replace the old version of Oasis Node with version 21.2.8 and restart your node.
Since Oasis Core 21.2.8 is otherwise compatible with the current consensus layer protocol, you may upgrade your Mainnet node to this version at any time.
For this upgrade, do NOT wipe state.
- Once reaching the designated upgrade epoch, your node will stop and needs to be upgraded to Oasis Core 21.2.8.
- If you upgraded your node to Oasis Core 21.2.8 before the upgrade epoch was reached, you only need to restart your node for the upgrade to proceed.
- Otherwise, you need to upgrade your node to Oasis Core 21.2.8 first and then restart it.
If you use a process manager like systemd or Supervisor, you can configure it to restart the Oasis Node automatically.
The Mainnet's genesis file and the genesis document's hash will remain the same.
BadgerDB v2 to v3 Migration
This upgrade will upgrade Oasis Core to version 21.2.x which includes the new BadgerDB v3.
Since BadgerDB's on-disk format changed in v3, it requires on-disk state migration. The migration process is done automatically and makes the following steps:
-
Upon startup, Oasis Node will start migrating all
<DATA-DIR>/**/*.badger.db
files (Badger v2 files) and start writing Badger v3 DB to files with the.migrate
suffix. -
If the migration fails in the middle, Oasis Node will delete all
<DATA-DIR>/**/*.badger.db.migrate
files the next time it starts and start the migration (of the remaining<DATA-DIR>/**/*.badger.db
files) again.
-
If the migration succeeds, Oasis Node will append the
.backup
suffix to all<DATA-DIR>/**/*.badger.db
files (Badger v2 files) and remove the.migrate
suffix from all<DATA-DIR>/**/*.badger.db.migrate
files (Badger v3 files).
The BadgerDB v2 to v3 migration is very I/O intensive (both IOPS and throughput) and may take a couple of hours to complete.
To follow its progress, run:
shopt -s globstar
du -h <DATA-DIR>/**/*.badger.db* | sort -h -r
and observe the sizes of various *.badger.db*
directories.
For example, if it outputted the following:
55G data/tendermint/data/blockstore.badger.db
37G data/tendermint/abci-state/mkvs_storage.badger.db.backup
32G data/tendermint/abci-state/mkvs_storage.badger.db
16G data/tendermint/data/blockstore.badger.db.migration
2.9G data/tendermint/data/state.badger.db
62M data/persistent-store.badger.db.backup
2.1M data/tendermint/abci-state/mkvs_storage.badger.db.backup/checkpoints
1.1M data/tendermint/abci-state/mkvs_storage.badger.db.backup/checkpoints/4767601/ca51b06a054b69f2c18b9781ea42f0b00900de199c1937398514331b0d136ec3/chunks
1.1M data/tendermint/abci-state/mkvs_storage.badger.db.backup/checkpoints/4767601/ca51b06a054b69f2c18b9781ea42f0b00900de199c1937398514331b0d136ec3
1.1M data/tendermint/abci-state/mkvs_storage.badger.db.backup/checkpoints/4767601
1.1M data/tendermint/abci-state/mkvs_storage.badger.db.backup/checkpoints/4757601/2ec3a28b1f4a2fcce503f2e80eb5d77b6c0a4d1075e8a14d880ac390338a855e/chunks
1.1M data/tendermint/abci-state/mkvs_storage.badger.db.backup/checkpoints/4757601/2ec3a28b1f4a2fcce503f2e80eb5d77b6c0a4d1075e8a14d880ac390338a855e
1.1M data/tendermint/abci-state/mkvs_storage.badger.db.backup/checkpoints/4757601
36K data/persistent-store.badger.db
20K data/tendermint/data/evidence.badger.db
then the mkvs_storage.badger.db
was already migrated:
- old BadgerDB v2 directory:
37G data/tendermint/abci-state/mkvs_storage.badger.db.backup
- new BadgerDB v3 directory:
32G data/tendermint/abci-state/mkvs_storage.badger.db
and now the blockstore.badger.db
is being migrated:
-
current BadgerDB v2 directory:
55G data/tendermint/data/blockstore.badger.db
-
new BadgerDB v3 directory:
16G data/tendermint/data/blockstore.badger.db.migration
Note that usually, the new BadgerDB v3 directory is smaller due to less fragmentation.
Extra storage requirements
Your node will thus need to have extra storage space to store both the old and the new BadgerDB files.
To see estimate how much extra space the migration will need, use the du
tool:
shopt -s globstar
du -h <DATA-DIR>/**/*.badger.db | sort -h -r
This is an example output from a Mainnet node that uses /srv/oasis/node
as the <DATA-DIR>
:
43G /srv/oasis/node/tendermint/data/blockstore.badger.db
28G /srv/oasis/node/tendermint/abci-state/mkvs_storage.badger.db
311M /srv/oasis/node/tendermint/data/state.badger.db
2.0M /srv/oasis/node/tendermint/abci-state/mkvs_storage.badger.db/checkpoints
996K /srv/oasis/node/tendermint/abci-state/mkvs_storage.badger.db/checkpoints/4517601
996K /srv/oasis/node/tendermint/abci-state/mkvs_storage.badger.db/checkpoints/4507601
992K /srv/oasis/node/tendermint/abci-state/mkvs_storage.badger.db/checkpoints/4517601/ba6218d7be2df31ba6e7201a8585c6435154728e55bbb7df1ffebe683bf60217
992K /srv/oasis/node/tendermint/abci-state/mkvs_storage.badger.db/checkpoints/4507601/1e0bf592bb0d99832b13ad91bc32aed018dfc2639e07b93a254a05f6791a19ac
984K /srv/oasis/node/tendermint/abci-state/mkvs_storage.badger.db/checkpoints/4517601/ba6218d7be2df31ba6e7201a8585c6435154728e55bbb7df1ffebe683bf60217/chunks
984K /srv/oasis/node/tendermint/abci-state/mkvs_storage.badger.db/checkpoints/4507601/1e0bf592bb0d99832b13ad91bc32aed018dfc2639e07b93a254a05f6791a19ac/chunks
148K /srv/oasis/node/persistent-store.badger.db
36K /srv/oasis/node/tendermint/data/evidence.badger.db
After you've confirmed your node is up and running, you can safely delete all the <DATA-DIR>/**/*.badger.db.backup
files.
Extra memory requirements
BadgerDB v2 to v3 migration can use a number of Go routines to migrate different database files in parallel.
However, this comes with a memory cost. For larger database files, it might need up to 4 GB of RAM per database, so we recommend lowering the number of Go routines BadgerDB uses during migration (badger.migrate.num_go_routines
) if your node has less than 8 GB of RAM.
If your node has less than 8 GB of RAM, set the number of Go routines BadgerDB uses during migration to 2 (default is 8) by adding the following to your node's config.yml
:
# BadgerDB configuration.
badger:
migrate:
# Set the number of Go routines BadgerDB uses during migration to 2 to lower
# the memory pressure during migration (at the expense of a longer migration
# time).
num_go_routines: 2
Installing Oasis App 2.3.1 to Your Ledger
This manual installation procedure is needed until the latest version of the Oasis App, version 2.3.1, becomes available through Ledger Live's Manager.
Unlike Nano S devices, Nano X devices are locked meaning one cannot manual install the latest version of the Oasis App on them. If you use a Nano X device, you will need to temporarily switch to a Nano S device or wait for the new version of the Oasis App to be available through Ledger Live's Manager.
Update Firmware to Version 2.0.0
First, make sure the firmware on your Nano S is up-to-date. At least version 2.0.0 released on May 4, 2021, is required. Follow Ledger's instructions for updating the firmware on your Nano S.
Install Prerequisites for Manual Installation
The manual installation process relies on some tooling that needs to be available on the system:
Most systems should already have Python pre-installed.
To install Python tools for Ledger Blue, Nano S and Nano X, use pip:
pip3 install --upgrade ledgerblue
You might want to install the packages to a Python virtual environment or via so-called User install (i.e. isolated to the current user).
Download Oasis App 2.3.1
Download the Oasis App 2.3.1 installer for Nano S from Zondax's Oasis App GitHub repo.
Install Oasis App 2.3.1
Make the downloaded installer executable by running:
chmod +x installer_s.sh
Connect you Nano S and unlock it. Then execute the installer:
./installer_s.sh load
Your Nano S will give you the option to either:
- Deny unsafe manager, or
- review the Public Key and Allow unsafe manager.
First review the public key and ensure it matches the Generated random root public key
displayed in the terminal.
Then double press the Allow unsafe manager option.
If there is an existing version of the Oasis App installed on your Nano S, you will be prompted with the Uninstall Oasis screen, followed by reviewing the Identifier (it will depend on the version of the Oasis App you have currently installed) and finally confirming deletion on the Confirm action screen.
After the new version of the Oasis App has finished loading, you will be prompted with the Install app Oasis screen, followed by reviewing the Version, Identifier and Code Identifier screens_._ Ensure the values are as follows:
- Version: 2.3.1
- Identifier (Application full hash on the terminal):
E0CB424D3B1C2A0F694BCB6E99C3B37C7685399D59DD12D7CF80AF4A487882B1
- Code Identifier:
C17EBE7CD356D01411A02A81C64CDA3E81F193BDA09BEBBD0AEAF75AD7EC35E3
Finally, confirm installation of the new app by double pressing on the Perform installation screen. Your Ledger device will ask for your PIN again.
Installing Oasis App 2.3.1 on a Nano S with the firmware version < 2.0.0 (e.g. 1.6.1) will NOT fail. It will show a different Identifier when installing the app which will NOT match the Application full hash shown on the terminal. However, opening the app will not work and it will "freeze" your Nano S device.
Verify Installation
Open the Oasis App on your Nano S and ensure the Version screen shows version 2.3.1.
Starting the manually installed version of the Oasis App will always show the This app is not genuine screen, followed by Identifier (which should match the Identifier value above) screen. Finally, open the application by double pressing on the Open application screen.
After you've signed your governance.CastVote
transaction, you can safely downgrade Oasis App to the latest official version available via Ledger Live, version 1.8.2.
To do that, just open Ledger Live's Manager and it will prompt you to install version 1.8.2.
2021-04-28 (16:00 UTC) - Cobalt Upgrade
- Upgrade height: upgrade is scheduled to happen at epoch 5046.
We expect the Mainnet network to reach this epoch at around 2021-04-28 12:00 UTC.
Instructions - Before upgrade
- Make sure you are running the latest Mainnet-compatible Oasis Node version: 20.12.7.
- If you are running a different 20.12.x Oasis Node version, update to version 20.12.7 before the upgrade.
Version 20.12.7 is backwards compatible with other 20.12.x releases, so upgrade can be performed at any time by stopping the node and replacing the binary.
-
To ensure your node will stop at epoch 5046 submit the following upgrade descriptor at any time before the upgrade:
{
"name": "mainnet-upgrade-2021-04-28",
"method": "internal",
"identifier": "mainnet-upgrade-2021-04-28",
"epoch": 5046
}
The upgrade descriptor contains a non-existing upgrade handler and will be used to coordinate the network shutdown, the rest of the upgrade is manual.
Runtime operators
Following section is relevant only for runtime operators that are running storage nodes for active runtimes on the Mainnet.
This upgrade requires a runtime storage node migration to be performed before the upgrade genesis is published. This can be done before the upgrade epoch is reached by stopping all runtime nodes and running the migration.
Backup your node's data directory
To prevent irrecoverable runtime storage data corruption/loss in case of a failed storage migration, backup your node's data directory.
For example, to backup the /serverdir/node
directory using the rsync tool, run:
rsync -a /serverdir/node/ /serverdir/node-BACKUP/
The storage database on all storage nodes needs to be migrated with the following command (using the 21.1.1 binary):
oasis-node storage migrate \
--datadir <NODE-DATADIR> \
--runtime.supported <RUNTIME-ID>
After the migration to v5 completes, you will see an output similar to:
...
- migrating from v4 to v5...
- migrating version 24468...
- migrated root state-root:195cf7a9a103e7300b2bb4e537cb9935cbebd83e448e67aa55433861a6ad7426 -> state-root:cea105a5d701deab935b94af9e8e0c5af5dcdb61c242bf434da9f11aa8d110ba
- migrated root io-root:0850c5a33ee7f45aa92724b7d5f28c9ac9ae8799b88cc5be9773e8aba9526ca7 -> io-root:19713a2b44e1bf868ebee43c36872baa3058870bb890a5e25d1c4cea2622be77
- migrated root io-root:477391131f60ac2c22bce9167c7e3783a13d4fb81fddd2d388b4ead6a586fe52 -> io-root:f29f86d491303c5fd7b3572e97cbd65b7487b6b4ac519623afd161cc2e4678b7
Take note of the displayed state-root
and report it to the Foundation, as it needs to be included in the upgrade's new genesis file. Keep the runtime nodes stopped until the upgrade epoch is reached. At upgrade epoch, upgrade the nodes by following the remaining steps above.
Instructions - Upgrade day
Following steps should be performed on 2021-04-28 only after the network has reached the upgrade epoch and has halted:
- Download the genesis file published in the Cobalt Upgrade release.
Mainnet state at epoch 5046 will be exported and migrated to a 21.1.x compatible genesis file. Upgrade genesis file will be published on the above link soon after reaching the upgrade epoch.
- Verify the provided Cobalt upgrade genesis file by comparing it to network state dump. See instructions in the Handling Network Upgrades guide.
- Replace the old genesis file with the new Cobalt upgrade genesis file.
- Stop your node (if you haven't stopped it already by submitting the upgrade descriptor).
- Replace the old version of Oasis Node with version 21.1.1.
- Wipe state.
- Start your node.
For more detailed instructions, see the Handling Network Upgrades guide.
Additional notes
Examine the Change Log of the 21.1.1 (and 21.0) releases.
Runtime operators
Note the following configuration change in the 21.0 release.
Storage access policy changes
Due to the changes in the default access policy on storage nodes, at least one of the storage nodes should be configured with the worker.storage.public_rpc.enabled
flag set to true
.
Otherwise, external runtime clients wont be able to connect to any storage nodes.
2020-11-18 (16:00 UTC) - Mainnet
- Block height when Mainnet Beta network stops: 702000.
We expect the Mainnet Beta network to reach this block height at around 2020-11-18 13:30 UTC.
- Upgrade window:
- Start: 2020-11-18T16:00:00Z.
- End: After nodes representing 2/3+ stake do the upgrade.
Instructions
- Download Oasis Node version 20.12.2, while continuing to run version 20.10.x.
- (optional) Use Oasis Node version 20.12.2 to dump network state at the specified block height. It will connect to the running version 20.10.x node.
- Download the Mainnet genesis file published in the 2020-11-18 release.
- (optional) Verify the provided Mainnet genesis file by comparing it to network state dump. See instructions in the Handling Network Upgrades guide.
- Replace the old Mainnet Beta genesis file with the Mainnet genesis file.
- Stop your node.
- Remove the old 20.10.x version of Oasis Node.
- Wipe state.
- Update your node's configuration per instructions in Configuration changes below.
- Start your node.
This time, we recommend dumping the network state with the upgraded Oasis Node binary so that the genesis file will be in the canonical form.
The canonical form will make it easier to compare the obtained genesis file with the one provided by us.
For more detailed instructions, see the Handling Network Upgrades guide.
Configuration changes
Since we are upgrading to the Mainnet, we recommend you change your node's configuration and disable pruning of the consensus' state by removing the consensus.tendermint.abci.prune
key.
For example, this configuration:
...
# Consensus backend.
consensus:
# Setting this to true will mean that the node you're deploying will attempt
# to register as a validator.
validator: true
# Tendermint backend configuration.
tendermint:
abci:
prune:
strategy: keep_n
# Keep ~7 days of data since block production is ~1 block every 6 seconds.
# (7*24*3600/6 = 100800)
num_kept: 100800
core:
listen_address: tcp://0.0.0.0:26656
...
Becomes:
...
# Consensus backend.
consensus:
# Setting this to true will mean that the node you're deploying will attempt
# to register as a validator.
validator: true
# Tendermint backend configuration.
tendermint:
core:
listen_address: tcp://0.0.0.0:26656
...
2020-10-01 - Mainnet Beta
Instructions
- Stop your node.
- Wipe state.
- Replace the old genesis file with the Mainnet Beta genesis file published in the 2020-10-01 release.
- Start your node.
You should keep using Oasis Core version 20.10.
For more detailed instructions, see the Handling Network Upgrades guide.
2020-09-22 - Mainnet Dry Run
Instructions
- This is the initial deployment.