GXChain Docs

# witness_node launch FAQ

# 1. How to start the witness_node program normally

It is recommended to start the program as follows (depending on your own situation, modify the parameter values ​​and port numbers as appropriate)

export LC_ALL=C
nohup ./programs/witness_node/witness_node --data-dir=trusted_node  --rpc-endpoint="0.0.0.0:28090" --p2p-endpoint="0.0.0.0:6789" --track-account "\"1.2.241993\"" --max-ops-per-account=-1  --fast-replay 1>nohup.out 2>&1 &

Start up as above, the console output will be redirected to nohup.out, if there is an exception, you may view the nohup.out location problem.

# 2. How to judge the witness_node is running normally

Observe the background log file trusted_node/logs/witness.log. If you continue to receive the block on the network, it means the node is running normally, as follows

2018-08-23T07:10:39 th_a:invoke handle_block         handle_block ] Got block: #6939202 time: 2018-08-23T07:10:39 latency: 29 ms from: init6  irreversible: 6939194 (-8)		application.cpp:487
2018-08-23T07:10:42 th_a:invoke handle_block         handle_block ] Got block: #6939203 time: 2018-08-23T07:10:42 latency: 30 ms from: init4  irreversible: 6939195 (-8)		application.cpp:487
2018-08-23T07:10:45 th_a:invoke handle_block         handle_block ] Got block: #6939204 time: 2018-08-23T07:10:45 latency: 32 ms from: init0  irreversible: 6939195 (-9)		application.cpp:487
2018-08-23T07:10:48 th_a:invoke handle_block         handle_block ] Got block: #6939205 time: 2018-08-23T07:10:48 latency: 34 ms from: init5  irreversible: 6939197 (-8)		application.cpp:487
2018-08-23T07:10:51 th_a:invoke handle_block         handle_block ] Got block: #6939206 time: 2018-08-23T07:10:51 latency: 34 ms from: init1  irreversible: 6939198 (-8)		application.cpp:487
2018-08-23T07:10:54 th_a:invoke handle_block         handle_block ] Got block: #6939207 time: 2018-08-23T07:10:54 latency: 36 ms from: bob  irreversible: 6939199 (-8)			application.cpp:487
2018-08-23T07:10:57 th_a:invoke handle_block         handle_block ] Got block: #6939208 time: 2018-08-23T07:10:57 latency: 37 ms from: init1  irreversible: 6939200 (-8)		application.cpp:487
2018-08-23T07:11:00 th_a:invoke handle_block         handle_block ] Got block: #6939209 time: 2018-08-23T07:11:00 latency: 39 ms from: init7  irreversible: 6939200 (-9)		application.cpp:487
2018-08-23T07:11:03 th_a:invoke handle_block         handle_block ] Got block: #6939210 time: 2018-08-23T07:11:03 latency: 41 ms from: bao  irreversible: 6939201 (-9)			application.cpp:487
2018-08-23T07:11:06 th_a:invoke handle_block         handle_block ] Got block: #6939211 time: 2018-08-23T07:11:06 latency: 41 ms from: init8  irreversible: 6939202 (-9)		application.cpp:487
2018-08-23T07:11:09 th_a:invoke handle_block         handle_block ] Got block: #6939212 time: 2018-08-23T07:11:09 latency: 44 ms from: init2  irreversible: 6939203 (-9)		application.cpp:487
2018-08-23T07:11:12 th_a:invoke handle_block         handle_block ] Got block: #6939213 time: 2018-08-23T07:11:12 latency: 45 ms from: init4  irreversible: 6939204 (-9)

# 3. Error when blocking a block

Observe the background log file trusted_node/logs/witness.log. If the log continues to report errors, such as `unlinkable block", "block does not link to known chain", this is a block synchronization error.

Solution:

  • First close the program, use --replay-blockchain to restart, wait for the startup to complete, observe whether the history.log log is normal.

  • Block sync exception, it is possible that the local blockchain file is broken. Need to stop the witness_node program, then delete the trusted_node directory and restart the witness_node.

  • Or you do not need to delete the trusted_node directory. The startup command plus the parameter --resync-blockchain will resynchronize the block.

# 4. How to properly shut down witness_node

Witness_node saves all data in the form of an object in memory. When the program exits normally, the data in the memory is written to the disk, so the process cannot be forced to be killed, otherwise the memory database will be broken.

  • If the witness_node is not running in the background, execute Ctrl + C once and then wait for the program to save the memory data and then exit automatically.

  • If the witness_node is running in the background, execute kill -s SIGINT $(pgrep witness_node), and wait for the program to save the memory data and then exit automatically. You can't use kill -9, otherwise the next time you start the index, the index will be rebuilt.

# 5. Witness_node process is normal, but does not receive blocks from the network

  • Check the background log file, found no error message, and the node did not receive a new block from the network. At this point, your version of witness_node may be low. Please visit the [github release page] (https://github.com/gxchain/gxb-core/releases/latest) to download the latest package.

  • You can view the duration_node version with the -v parameter:

$./programs/witness_node/witness_node -v
Version: 1.0.180713-379-g4c9a2f4
SHA: 4c9a2f4e168503abe3ce1432c699f88b8babe356
Timestamp: 4 hours ago
SSL: OpenSSL 1.0.2o  27 Mar 2018
Boost: 1.67
Websocket++: 0.7.0

# 6.What operating system does BP use, and what are the requirements for hardware including bandwidth?

  • ubuntu 14.04 64 os
  • 4 core 32G memory 200G hard disk, 50MB+ bandwidth, according to the flow rate, double machine disaster recovery
  • Can use cloud server

# 7. When the witness_node starts, the port error is reported.

Check if multiple witness_nodes are started on the server, causing port conflicts.

ps xf | grep witness_node

If you have started multiple witness_nodes, you need to close them all and restart them:

for i in $(pgrep witness_node); do echo kill -9 $i; done

# 8. Witness_node failed to start, no log, process does not exist

Modify the local config.ini file as follows:

[logger.default]
level=debug
appenders=stderr,FILE

Then restart, the error message will be printed to the console