Hyperledger Fabric - How to clear out the DEV environment after each blockchain network test? - crash

Question

Fabric 1.4.3 version. Blockchain network with 1 Oderer (solo) + 1 Org, running on Docker.
Trying to instantiate chaincode, due to PANIC error on PEER0, and peer crashes.
Impossible to instantiate chaincode, because PEER0 crashes doing the process.
At CLI docker prompt, I did this command sequence:
1) $> peer channel create -o $ORDERERNAME -c $CHANNELNAME -f $CONFIGTXFOLDER/devchannel.tx --tls --cafile=$ORDERER_TLSCACERT
Result in Cli: UTC [cli.common] readBlock -> INFO 04e Received block: 0
2) $> peer channel join -o $ORDERERNAME -b $CONFIGTXFOLDER/devgenesis.block --tls --cafile=$ORDERER_TLSCACERT
Result in Cli: UTC [channelCmd] executeJoin -> INFO 03e Successfully submitted proposal to join channel
3) $> peer chaincode install -n $CHCODENAME -p $CHCODEPATH -v $CHCODEVERSION -l node --tls --cafile $ADMIN_PEER_TLSCACERT
Result in Cli: UTC [chaincodeCmd] install -> INFO 04a Installed remotely response:<status:200 payload:"OK" >
4) $> peer chaincode instantiate -C $CHANNELNAME -n $CHCODENAME -v $CHCODEVERSION -o $ORDERERNAME -c '{"Args":["init","a","100","b","200"]}' -P "AND ('GuaraniMSP.admin')" --tls --cafile $ORDERER_TLSCACERT --tlsRootCertFiles $CORE_PEER_TLS_ROOTCERT_FILE
Result before crash in PEER0:
UTC [gossip.state] commitBlock -> ERRO 87e Got error while committing(unexpected Previous block hash. Expected PreviousHash = [c87a4b77e4c790f78b0c2e3c97d97de9907a09daf5dc2f039c7e3b3e1440f5d1], PreviousHash referred in the latest block= [953e31164a84d6d1b9b446130d1e7d5af8ede818284e8fa7c315b2125b519e38]
github.com/hyperledger/fabric/common/ledger/blkstorage/fsblkstorage.(*blockfileMgr).addBlock
[...]
UTC [gossip.state] deliverPayloads -> PANI 87f Cannot commit block to the ledger due to unexpected Previous block hash. Expected PreviousHash = [c87a4b77e4c790f78b0c2e3c97d97de9907a09daf5dc2f039c7e3b3e1440f5d1], PreviousHash referred in the latest block= [953e31164a84d6d1b9b446130d1e7d5af8ede818284e8fa7c315b2125b519e38]
github.com/hyperledger/fabric/common/ledger/blkstorage/fsblkstorage.(*blockfileMgr).addBlock
[...]
/opt/go/src/runtime/asm_amd64.s:1333
panic: Cannot commit block to the ledger due to unexpected Previous block hash. Expected PreviousHash = [c87a4b77e4c790f78b0c2e3c97d97de9907a09daf5dc2f039c7e3b3e1440f5d1], PreviousHash referred in the latest block= [953e31164a84d6d1b9b446130d1e7d5af8ede818284e8fa7c315b2125b519e38]
github.com/hyperledger/fabric/common/ledger/blkstorage/fsblkstorage.(*blockfileMgr).addBlock
[...]
I opened an issue at Hyperledger Fabric JIRA about this situation, and I received the information that I need to clear out the environment to ensure there are no artifacts from prior trials.
Issue FABB-147 at Hyperledger Fabric (My logs are there).
I saw a problem documented close to my, but with differences, at peers get crash after anchor peer update
Now I have to certify that my Hyperledger Fabric environment is clear. But how?
Is there any documented procedure or checklist to verify if there are no artifacts from prior trials in Hyperledger Fabric?
If I suppress the ORDERER_GENERAL_GENESISFILE and ORDERER_GENERAL_GENESISPROFILE settings for orderer.yaml and docker-compose.yaml, will the orderer element start normally, and then the network will understand that it should use the genesis.block informed when creating the new channel?
Thanks in advance for your help.

Answer 1

Run this command docker system prune
Start Your network
again
hope this help u as it helped me

Answer 2

While not an answer to cleaning your environment, here's a possible solution to your underlying problem.
The peer channel create command will generate a $CHANNELNAME.block file in its current working directory. This is the block you need to use when running peer channel join, not a genesis block.
$> peer channel create ... -c $CHANNELNAME
$> peer channel join ... -b $CONFIGTXFOLDER/devgenesis.block
Will ruin your day by causing your peers to crash most of the time, and rarely work after recreating the ledger from a new genesis block. Try this instead:
$> peer channel create ... -c $CHANNELNAME
$> peer channel join ... -b $CHANNELNAME.block

Answer 3

I was using the genesis block from the system fabric network, loaded on the Ordener (from orderer.yaml or docker-compose.yaml), to join the channel, and that was wrong.
The source of my error was that I was using the system genesis block instead of using the genesis block generated in the peer channel create command.
I understood that the peer channel create command uses a transaction block model (.tx) as input, and as an output it generates a genesis block (.block) that must be used as input in the peer channel join command.
The correct sequence is:
Step 1) $> peer channel create -o $ORDERERNAME -c $CHANNELNAME -f $CONFIGTXFOLDER/devchanneltrack.tx --outputBlock $CHANNELFOLDER/devchannelgen.block --tls --cafile=$ORDERER_TLSCACERT
Step 2) $> peer channel join -o $ORDERERNAME -b $CHANNELFOLDER/devchannelgen.block --tls --cafile=$ORDERER_TLSCACERT
I made corrections to the scripts, and now I can join the channel. Then I did the installation and instantiation of new chaincodes on the channel, and they worked perfectly.