Closed dkaparis closed 4 years ago
Please try to do chown -R 777 data
from your mn-bootstrap dir and restart node.
Please try to do
chown -R 777 data
from your mn-bootstrap dir and restart node.
I suppose you mean chmod -R 777 data
This appears to solve the issue - all containers stay up without restarting so far. Perhaps it should be added to the documentation.
Oh, sorry, my bad.
Perhaps it should be added to the documentation.
@strophy Could you update the README, please?
I cannot reproduce this bug after 5-6 attempts following different sequences setting up from scratch. I think making the datadir globally writeable is a hack, I would rather find out how we are ending up in this state, since two separate users have come forward with this problem now. It's most likely a missing step during setup, @dkaparis can you verify you added your non-root user to the docker
group and refreshed your environment as described here?
... @dkaparis can you verify you added your non-root user to the
docker
group and refreshed your environment as described here?
I ran the whole setup and do the starting from the root account, so I didn't. Maybe that's where the problem lies, if so it should be noted. As it is now, these post-installation steps are described as optional.
The post-installation steps are only optional if you do not use the CLI. The CLI must be run as a non-root user, or the container will not be able to write to the filesystem. I think I have tracked down the error, you need to make sure that the npm install
command in particular is NOT run as the root user, i.e. do not use sudo. I worked through this with another user experiencing the same problem and we narrowed the bare minimum of steps for a clean install under Ubuntu 20.04 LTS, starting as root:
apt update
apt install docker.io docker-compose nodejs npm
adduser strophy
usermod -aG sudo,docker strophy
su - strophy
git clone https://github.com/dashevo/mn-bootstrap
cd mn-bootstrap
nano configs/evonet/core/dashd.conf //temporary workaround, bump evonet version to evonet-5
npm install
sudo npm link
mn start evonet <ip> 19999 -p <bls-privkey>
Can you please try this and confirm it works? Then I'll look into making the documentation more clear.
The post-installation steps are only optional if you do not use the CLI. The CLI must be run as a non-root user, or the container will not be able to write to the filesystem. I think I have tracked down the error, you need to make sure that the
npm install
command in particular is NOT run as the root user, i.e. do not use sudo. I worked through this with another user experiencing the same problem and we narrowed the bare minimum of steps for a clean install under Ubuntu 20.04 LTS, starting as root:apt update apt install docker.io docker-compose nodejs npm adduser strophy usermod -aG sudo,docker strophy su - strophy git clone https://github.com/dashevo/mn-bootstrap cd mn-bootstrap nano configs/evonet/core/dashd.conf //temporary workaround, bump evonet version to evonet-5 npm install sudo npm link mn start evonet <ip> 19999 -p <bls-privkey>
Can you please try this and confirm it works? Then I'll look into making the documentation more clear.
Following the above procedure from a clean install (modified for Debian 10), all containers run except for dash_masternode_evonet_drive_tendermint_1 which is again restarting.
docker logs dash_masternode_evonet_drive_tendermint_1
...
goroutine 1 [running]:
github.com/tendermint/tendermint/privval.FilePVKey.Save(0xc000f228e0, 0x14, 0x20, 0x11c4440, 0xc000f22900, 0x11c4400, 0xc0000f5c80, 0xc0002ec760, 0x1d)
/home/archie/Documents/tendermint/privval/file.go:61 +0x182
github.com/tendermint/tendermint/privval.(*FilePV).Save(0xc0002e83c0)
/home/archie/Documents/tendermint/privval/file.go:262 +0xa1
github.com/tendermint/tendermint/cmd/tendermint/commands.initFilesWithConfig(0xc000f04000, 0x0, 0x0)
/home/archie/Documents/tendermint/cmd/tendermint/commands/init.go:37 +0xbf6
github.com/tendermint/tendermint/cmd/tendermint/commands.initFiles(0x187d0a0, 0x18a7a48, 0x0, 0x0, 0x0, 0x0)
/home/archie/Documents/tendermint/cmd/tendermint/commands/init.go:23 +0x2d
github.com/spf13/cobra.(*Command).execute(0x187d0a0, 0x18a7a48, 0x0, 0x0, 0x187d0a0, 0x18a7a48)
/home/archie/go/pkg/mod/github.com/spf13/cobra@v0.0.1/command.go:698 +0x42c
github.com/spf13/cobra.(*Command).ExecuteC(0x187e3c0, 0x2, 0xc00000cbc0, 0xf47739)
/home/archie/go/pkg/mod/github.com/spf13/cobra@v0.0.1/command.go:783 +0x2c9
github.com/spf13/cobra.(*Command).Execute(...)
/home/archie/go/pkg/mod/github.com/spf13/cobra@v0.0.1/command.go:736
github.com/tendermint/tendermint/libs/cli.Executor.Execute(0x187e3c0, 0x10970b8, 0x2, 0xc000135ba0)
/home/archie/Documents/tendermint/libs/cli/setup.go:89 +0x3c
main.main()
/home/archie/Documents/tendermint/cmd/tendermint/main.go:45 +0x248
panic: open /data/write-file-atomic-246361497058367640: permission denied
This one is related #78
Is it still actual, since we moved data to docker volumes?
Right, I'm fairly sure this can be closed because docker volumes handle permissions much better than bind mounting the datadir. @dkaparis please open another issue if you are still having problems.
Expected Behavior
Docker containers dashmasternode* should stay up
Current Behavior
Containers
dash_masternode_evonet_drive_tendermint_1
,dash_masternode_evonet_drive_abci_1
anddash_masternode_evonet_core_1
are continuously restartingPartial log outputs:
docker logs dash_masternode_evonet_drive_tendermint_1
:docker logs dash_masternode_evonet_drive_abci_1
:docker logs dash_masternode_evonet_core_1
:Not sure if related, but in system log I see repeated entries like these:
Steps to Reproduce
mn start evonet <my-ip-addr> 9999 -p <operator-key>
Context
Although the masternode appears to be enabled and running when queried with the DashMasternodeTool, after some time status goes to POSE_BANNED
Your Environment
4.19.0-9-amd64 #1 SMP Debian 4.19.118-2+deb10u1 (2020-06-07) x86_64 GNU/Linux