ooibc88 / blockbench Goto Github PK

Hi guys,
I'm attempting to use Blockbench as part of my final third year university project, however, i'm having trouble configuring it. I have run the install (built peer), built the macro drivers, installed the micro dependencies via npm, patched, built and installed restclient-cpp. Now I am trying to run start-root and start-peer to try and actually run some benchmarks and this is where I run into trouble.

env.sh
https://pastebin.com/CiGZeSU5

hl_log (as created by start-root)
https://pastebin.com/dY4cJtSq

If someone could provide some help it would be greatly appreciated.

When I run the parity and ethereum , I meet the error “ driver: evm_utils.cc:334: std::__cxx11::string BBUtils::EVMUtils::lookup_smart_contract_address_or_die(const string&, const string&): Assertion `r.find(""result":null") == std::string::npos' failed. " . I found the reason is the code " assert(r.find(""result":null") == std::string::npos); " in function
"lookup_smart_contract_address_or_die" , I have print r，the information is " {"jsonrpc":"2.0","result":null,"id":1} " , but in parity ，the line of code is annotated ，so there haven't this error , so I want to know , whether is correct or not of the r .

I used two virtual machines to do this experiment, one as the host and one as the client. I have made the following changes:
I installed geth on both vms, and used ssh to connect two vms. I modified env.sh according to my own path and modified hosts and clients according to the ip address.All the dinhtta in the script is also changed to mine
What else do I need to do?

I am using Ubuntu 14.04 LTS and Rasberry Pi 2 for setting up the blockbench environment. I started with ethereum. After making changes in env.sh I start benchmark with ./run_bench 1 1 1.

The output from running the benchmark is at
https://pastebin.com/W6TKVS1Z

My env.sh is as follows
ETH_HOME=/home/sonam/blockbench/benchmark/ethereum
HOSTS=$ETH_HOME/hosts
CLIENTS=$ETH_HOME/clients
ETH_DATA=/home/sonam/data/eth
#LOG_DIR=$ETH_HOME/donothing_donothingresults_1
LOG_DIR=$ETH_HOME/yscb
EXE_HOME=/home/pi/blockbench/src/macro/kvstore
BENCHMARK=ycsb

My CustomGenesis file is as follows
{
"nonce": "0x0000000000000042",
"timestamp": "0x00",
"parentHash": "0x0000000000000000000000000000000000000000000000000000000000000000",
"extraData": "0x00",
"gasLimit": "0x8000000",
"difficulty": "0x1fffff",
"mixhash": "0x0000000000000000000000000000000000000000000000000000000000000000",
"coinbase": "0x3333333333333333333333333333333333333333",
"alloc":{ },
"config": {
"chainId": 9119,
"homesteadBlock": 0,
"eip155Block": 0,
"eip158Block": 0
}
}

Can you tell me where am going wrong?

Hi team,

I’m attempting to use blockbench as a basis of blockchain experimental work. When I tried single-node experiment running YSBC in local machine, while it seemed to work fine, logs showed something very suspicious. For example, printout messages and log files showed below, probably indicating that drivers and peers were not initialized properly. I just wonder if there’s any way to approach troubleshooting easily and what the right parameter is to test single-machine testing.

Thanks,
MJ

when it call the function "lookup_smart_contract_address_or_die" , it always return {"jsonrpc":"2.0","id":1,"result":null} , and at the last step , the value of receipt is not null , I print the value is "receipt: 0xad1deef543e1ced521bac1ffa1d4af629a1c32d3c34fc204fb0f58d513f5bcbf " , but when I run parity , it can successfully get the value of contractAddress by call the function "lookup_smart_contract_address_or_die" .

Hi, since you are using Hyperledger 0.6. with pbft I want to try to run a benchmark on other HL versions with kafka. I made some changes to env.sh, some start shells. After ./run-bench.sh I get this output and no data as outcome. Have you already made a benchmark with kafka?

peer: no process found
java: no process found
driver: no process found
peer: no process found
java: no process found
driver: no process found
driver: no process found
peer: no process found
java: no process found
driver: no process found
peer: no process found
java: no process found
hl_consensus_2.yaml 100% 2788 2.7KB/s 00:00
hl_consensus_2.yaml 100% 2788 2.7KB/s 00:00
bpeer = 192.168.0.50
192.168.0.48 index 0
IN START_CLIENTS 4 0 2 10
peer: no process found
java: no process found
driver: no process found
peer: no process found
java: no process found
driver: no process found
driver: no process found
peer: no process found
java: no process found
driver: no process found
peer: no process found
java: no process found

I found there have value " : " in pendingtx , so I look over the code " (*pendingtx_)[txn_hash] = start_time; " .
I print the txn_hash , and found the following problem :

I randomly choose from_addresses_ to deploy smart contracts and send transactions , but in this image , you can find the txn_hash is null in somtimes by same address . I can't understand it.

When I run ./run-bench.sh
Produced the following result：
java(18034): 不允许的操作（Not allowed）
java(18329): 不允许的操作
java(18621): 不允许的操作
java(19018): 不允许的操作
java(19197): 不允许的操作
java(51841): 不允许的操作
peer：没有发现操作（No operations found）
java：没有发现操作
driver：没有发现操作

I have configured global variables, the path should be no problem,Where did the problem occur?thanks!

What is the license of this project?
Could you add a LICENSE file to the root of the repository?

I am using 8 vms (with latest Ubuntu version) from google cloud and i am trying to set up the blockbench enviroment. I started with ethereum.
After the necessary changes i am trying to start the benchmark with ./run_bench 8 16 8
and in the log dir donothing_donothingresults_1 i get 3 files like that client_10.128.0.4_
and inside ith says ./driver: error while loading shared libraries: librestclient-cpp.so.1: cannot open shared object file: No such file or directory

here is the out put from running the bech

https://pastebin.com/TSCAT5ye

So it must be something with the ycsb right?

I want to run start_client.sh in the directory of blockbench/benchmark/parity/ , there have "nohup ../../src/ycsb/driver -db parity -threads $THREADS -P ../../src/ycsb/workloads/workloada.spec -endpoint $HOST -txrate $RATE > $LOG 2>&1 &" in the file ,but in the code that I download , don't have the ycsb directory ,only have src/macro/kvstore and src/macro/smallbank , do I need download other code in somewhere ??? Or install other things？？

@ug93tad @dloghin excuse me.
I am doing macro inspection of two nodes, I started from Ethereum. i run ./run-bench.sh 1 1 1 10 and i got result as follow:

according to start-multi-clients.sh ,i think i should got 192.168.43.208 index 0,but i got 192.168.43.208 index 1,and even more strange is ，I did not generate LOG diretory and no run echo IN START_CLIENTS(in start-client）
and for driver ：no process found ,i don't know how to deal with!
thank you!

Hi,

I could run some macro experiments. After evaluating the results I saw, that the driver executes only on one machine. My setup is starting from my local machine, use 2 VMs as Server and 2 VMs as Clients. My env.sh is set up correctly and I do run-bench.sh 2 4 4 %txrate. I also changed the conditions in the scripts to enter the loops, but got no success. Do you encountered the same problem?

i want to do a preliminary simulation ,two nodes.first install the geth ,then use genesis-block to bulid private blockchain, deploy smart contracts,then install the driver of kv_store,finally run_bench.sh. is the process right ?

At what point am I supposed to start the driver? Before or after i run the run-bench.sh 1 1 1??

What is the output of run-bench.sh supposed to look at?

It's a bit tiring that the documentation is not complete, and that the env.sh file is not the only file you have to edit before things start to work. Wish this was described in the documentation.

The project is very difficult to configure (indeed I am having a very hard time): it is due on the one hand to the repetition of environment declarations and on the other hand to the plenty of hard-coded configuration variables (e.g., dinhtta repeated in all files).

Therefore, it would be very helpful to adhere to the DRY principle and therefore avoid repetition and redeclaration of variables and hardcoded stuff. I would suggest using only a single configuration file in which all configuration variables are defined or even multiple files and a standard format, e.g. json, and using a command line tool, such as jq to read the configuration data.

Please provide the step by step procedure and resources required to run Blockbench on Ubuntu 16.04

Hello there,

I'm trying to run the analytics benchmarks on an ethereum private network however when I try to look through blocks the benchmarks will tell me there aren't 10 000 blocks minted.

When I check how many blocks there are on my geth console it reports there is indeed over 10 000 blocks

Any help would be appreciated

Thank You

Hi

In your hyperledger analytic contract Q2 doesn't seem to do what you say it does.

The Q2 is defined as

Compute the largest transaction value involving a given account between block i and block j.

However in Your bench_q2.js the method that you call is QueryAccount however this only returns the balance at a certain block. It doesn't look through blocks i to j and it doesn't return the max transaction value

RangeAccount still just returns a balance

In your paper the code you provided was

I'm just wondering where and how to implement Q2

Hello I'm trying to run the second analytic query for Hyperledger

The description for Q2 is

Q2: Compute the largest transaction value involving a given account between block i and block j

However when looking at the code for bench_q2.js it calls the QueryAccount function

The QueryAccount function only takes in two arguments, Account number and Block Number. Which is not what Q2 describes

When I change the function to RangeAccount, which a function that takes the correct arguments, the only answer every returned is -1 which means it's not searching through blocks i to j

Any help would be appreciated

Thank You

The paper describes that one needs to implement a certain interface called as IBlockchainConnector. However I could not find it anywhere in the code. How would I go about doing this?

When I run start-miner.sh , the error message is "flag provided but not defined: -voteaccout " in log file.

the problem command is " geth --datadir=~/chain-data-quorum_vote --rpccorsdomain "*" --nodiscover --voteaccout "0xeb6902c98ed64b7b16f743a7c09b36b8fa1aff59" --votepassword "" --rpc --rpcaddr 0.0.0.0 --rpcport 8000 "

So when i was creating an entity and lets say i wanted to use the waterflow image for one of the pieces, it auto applied it to all 3 cubes even when clicked on the one cube and no scroll bar to go choose width, or etc. So I closed and reload the program, now with nothing clicked when I import the textures, it replaces one texture instead of adding it. I'm going to downgrade but it would be nice for it to not replace textures.

If you close this with no info or to why like last time you guarentee you dont want users to use your program anymore. You used to be nice, show if i did something wrong, now you close an issue that is still prominent in your program with no aid, and your website doesnt let us download 1.9.1, so in other words please fix this or explain what Im doing wrong instead of closing with no help at all, its rude and mean and certifies i'll go use your competitors program which i dont want to do.

##I running the Ethereum benchmark for kvstore and return the zero result, why? How to solve??

Hi when I try to create the driver for kvstore I get his error

//usr/local/lib/librestclient-cpp.so: undefined reference to curl_global_cleanup@CURL_OPENSSL_3' //usr/local/lib/librestclient-cpp.so: undefined reference to curl_easy_reset@CURL_OPENSSL_3'
//usr/local/lib/librestclient-cpp.so: undefined reference to curl_easy_cleanup@CURL_OPENSSL_3' //usr/local/lib/librestclient-cpp.so: undefined reference to curl_easy_perform@CURL_OPENSSL_3'
//usr/local/lib/librestclient-cpp.so: undefined reference to curl_slist_free_all@CURL_OPENSSL_3' //usr/local/lib/librestclient-cpp.so: undefined reference to curl_easy_setopt@CURL_OPENSSL_3'
//usr/local/lib/librestclient-cpp.so: undefined reference to curl_easy_strerror@CURL_OPENSSL_3' //usr/local/lib/librestclient-cpp.so: undefined reference to curl_easy_init@CURL_OPENSSL_3'
//usr/local/lib/librestclient-cpp.so: undefined reference to curl_easy_getinfo@CURL_OPENSSL_3' //usr/local/lib/librestclient-cpp.so: undefined reference to curl_slist_append@CURL_OPENSSL_3'
//usr/local/lib/librestclient-cpp.so: undefined reference to `curl_global_init@CURL_OPENSSL_3'
collect2: error: ld returned 1 exit status
Makefile:15: recipe for target 'driver' failed
make: *** [driver] Error 1

I have libcurl installed so I don't know what's wrong

The Solidity code used to generate the bytecode here is not included in the repository.

The array in cpuheavy contract(blockbench-master/benchmark/contracts/ethereum/cpuheavy.sol)
is initialized in descending order. And in function quicksort(),i and j are initialized after the if condition.So it will always return the descending ordered array.How could it enter the while loop since the value of i and j will be always 0.

function quickSort(uint[] arr, uint left, uint right) internal {
if (i == j) return;
uint i = left;
uint j = right;

    uint pivot = arr[left + (right - left) / 2];
    while (i <= j) {
        while (arr[i] < pivot) i++;
        while (pivot < arr[j]) j--;
        if (i <= j) {
            (arr[i], arr[j]) = (arr[j], arr[i]);
            i++;
            j--;
        }
    }
    if (left < j)
        quickSort(arr, left, j);
    if (i < right)
        quickSort(arr, i, right);
}

}

when I print all txs , I found the value of "from" always is "0x0956e52cb729097e58fc3d945b0464e0ad68df2e" , I found the adress in config.toml.template . If all transactions are initiated by the same account, I think maybe it is wrong . And I query all accounts by call the api “eth_accounts” , it returns 32 accounts , but when I query their balance by call the api " eth_getBalance " , I found the value is "0x0" of all accounts . I don't know how to make all accounts have the opportunity to send transactions and has some balance ???

Hi there, I am not seeing the solidity contract source in smallbank.sol, rather is having KVstore source in it.

ref: blockbench/benchmark/contracts/ethereum/smallbank.sol
Would it be possible to update or share source of smallbank.sol please. Thank you

I am using Ubuntu 16.04.3 desktop 64-bit to run two virtual machines to test blockbench.

I am trying to run ./run-bench 1 1 1 1, but i get this output:

https://pastebin.com/EC0JbWVR

And these are my variables in the env.sh file:

ETH_HOME=/home/thesis/Documents/blockbench-master/benchmark/ethereum
HOSTS=$ETH_HOME/hosts
CLIENTS=$ETH_HOME/clients
ETH_DATA=/home/thesis/Documents/data
LOG_DIR=$ETH_HOME/donothing_donothingresults_1
EXE_HOME=/home/thesis/Documents/blockbench-master/src/smallbank/ycsb
BENCHMARK=ycsb

##comment these out for smallbank
#EXE_HOME=$ETH_HOME/../../src/smallbank/ethereum/ycsb
#BENCHMARK=smallbank

I tried cloning the CustomGenesis.json to create a CustomGenesis_1.json file, but I still got the same output (invalid genesis file .... ), and I tried following the advice on this page, but to no help.

What am I doing wrong? What can I do?

So when i was creating an entity and lets say i wanted to use the waterflow image for one of the pieces, it auto applied it to all 3 cubes even when clicked on the one cube and no scroll bar to go choose width, or etc. So I closed and reload the program, now with nothing clicked when I import the textures, it replaces one texture instead of adding it. I'm going to downgrade but it would be nice for it to not replace textures.

It is mentioned in the below pdf that ErisDB is under development,Is the implementation available now?
http://www.comp.nus.edu.sg/~ooibc/blockbench.pdf

When I look over the function poll_txs_by_block_number and poll_txs_by_block_hash ，I think that the return value of these two functions should be the same . But actually , in the code of " std::vectorstd::string ret = get_list_field(r, "transactions"); " , it can't get the transcation list , because in the function of get_list_field ,the string is split by ',' . I look over the return value of "eth_getBlockByNumber" API , the return transactions are split by ' } ' . I don't know if my conclusion is correct .

Another question in the code of
"
inline std::vectorstd::string poll_txs_by_block_number(const std::string &endpoint, int block_number) {
std::string request = GET_BLOCK_BY_NUMBER_PREFIX + std::to_string(block_number) +
GET_BLOCK_BY_NUMBER_SUFFIX;
auto r = send_jsonrpc_request(endpoint, REQUEST_HEADERS, request);
std::vectorstd::string ret = get_list_field(r, "transactions");
std::vectorstd::string uncles = get_list_field(r, "uncles");
for (std::string uncle : uncles) {
std::vectorstd::string uncletxs = poll_txs_by_block_hash(endpoint, uncle);
for (std::string tx : uncletxs) ret.push_back(tx);
}
return ret;
"
I can't understanding why do we need to add the return tx of function "poll_txs_by_block_hash" to the result ret of "get_list_field(r, "transactions") " , I think this will lead to repeated transcation in the vector .

blockbench/benchmark/ethereum/start-clients.sh
In the code line:

nohup ./driver -db ethereum -threads $1 -P workloads/workloada.spec -endpoint $host:8000 -txrate $4 -wt 60 > $LOG_DIR/client_$host"_"$1 2>&1 &

So how to build ./driver binary? Also go to blockbench/src/macro/kvstore then make command but this raise the issue:

core/hyperledger_utils.o: In function BBUtils::HLUtils::get_tip_block_number(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)': /opt/blockbench/src/macro/kvstore/core/hyperledger_utils.cc:143: undefined reference to RestClient::get(std::__cxx11::basic_string<char, std::char_traits, std::allocator > const&)'
core/hyperledger_utils.o: In function BBUtils::HLUtils::poll_txs_by_block_number(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, int)': /opt/blockbench/src/macro/kvstore/core/hyperledger_utils.cc:164: undefined reference to RestClient::get(std::__cxx11::basic_string<char, std::char_traits, std::allocator > const&)'
core/hyperledger_utils.o: In function BBUtils::send_jsonrpc_request(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)': /opt/blockbench/src/macro/kvstore/core/bb_utils.h:93: undefined reference to RestClient::post(std::__cxx11::basic_string<char, std::char_traits, std::allocator > const&, std::__cxx11::basic_string<char, std::char_traits, std::allocator > const&, std::__cxx11::basic_string<char, std::char_traits, std::allocator > const&)'
collect2: error: ld returned 1 exit status

So how to deal with that?

Thanks

Question about smallbank driver

I'm trying to run benchmark using Smallbank workload,But I'm not sure which one is right.
I saw the comment below:

And I tried to use the driver in src/macro/smallbank as document say, but I got a all zero results,like

In the last 2s, tx count = 0 latency = 0 outstanding request = 0
In the last 2s, tx count = 0 latency = 0 outstanding request = 395
In the last 2s, tx count = 0 latency = 0 outstanding request = 783
In the last 2s, tx count = 0 latency = 0 outstanding request = 1169
...

and tx count and latency are always 0, only the outstanding request increase. I'm still not sure about why I got this results

Question about code

Another question is about the true in the code below:

For Ethereum, smallbank is simulated via ycsb

In this way I can get some results, but the throughput I calculate is very large, like 10 times larger than YCSB workload. Is this result reasonable?

Hope to get help!

Hi, I keep getting node refused whenever I try to start run-bench.sh. What's going on?

The src/macro/kvstore folder contains two python script that look like they should help in parsing and aggregating the benchmark result from multiple clients. However, the script are hardly documented and also look for a "Throughput" string (line 31 in aggregate.py) in the log which is not printed by the driver.

Which logs are supposed to be processed with the scripts?

I can run experiments with 1 and 2 servers. But when I start with 3 servers, the client can't find an address, but the driver works. Do you have a suggestion why this appears. See Pastebin for log and client files. Server (just one) https://pastebin.com/Q49YcyVD and client https://pastebin.com/YpJPxauF

I have a Quorum private chain of 4 nodes spread across aws and azure instances. Can you please give me a brief overview of the setup I have to do for blockbench in this regards ?

Cheers
S

when I print the return information of function submit_writeCheck_txn , I found some account aren't unlock , but in the funciton WriteCheck ，the account is unlock by call unlock_address() before call submit_writeCheck_txn , so I can't understand why this error occurs .

Wrong GitHub, sorry. XD

@ijingo when I run node deploy.js there is an error encountered parse error, the specific information of this issue is listed as follows:

{"jsonrpc":"2.0","method":"deploy","params":{"type":1,"chaincodeID":{"path":"https://github.com/ijingo/chaincode-test/io"},"ctorMsg":{"args":["init"]}},"id":1}
problem with request: Parse Error

I am wondering how to solve it, is it related to the version of node? Or there is some mistakes in the code deploy.js?

when I print the value of endpoint , in parity,the value is 192.168.0.128:8000 , but in ethereum , the value is
192.168.0.128:8000/chaincode , caused an error in the function "lookup_smart_contract_address_or_die" . Should I modify the value of endpoint ???

How to change the configuration settings in env.sh ,for running Blockbench in Ethereum ?

I am using Ubuntu 16.04 LTS . After making changes in env.sh I run install.sh I am getting the error

Makefile:59: *** "No go in PATH: Check dependencies". Stop.

I even set GOPATH in .profile

export GOROOT=/opt/go
export PATH=$PATH:$GOROOT/bin
export GOPATH=$HOME/data/hyperledger/go
export PATH=$PATH:$GOPATH/bin

Can you tell me whether I am setting the path correct or not?

Also my env.sh is as follows:
HL_HOME=/home/user/blockbench/benchmark/hyperledger
HOSTS=$HL_HOME/hosts
CLIENTS=$HL_HOME/clients
HL_DATA=/home/user/data/hyperledger
HL_SOURCE=$HL_DATA/src/github.com/hyperledger/fabric
#LOG_DIR=$HL_HOME/donothing_donothingresults_1
LOG_DIR=$ETH_HOME/yscb
EXE_HOME=/home/user/blockbench/src/macro/kvstore
BENCHMARK=ycsb

Hello I'm trying to run the analytics benchmark on Hyperledger fabric v0.60.

I can deploy and populate the chain just fine but when I attempt to query it using bench_q1.js I receive this error

This is what shows up on my peer log

Any help would be great

Thank You

I found the EXE_HOME only is ".../src/macro/kvstore" and haven't such as "$ETH_HOME/../src/smallbank/ycsb" when I want start_client by smallbank . The directory of ".../src/macro/smallbank" don't have ycsb folder . I want to know hyperledger whether achieves smallbank or not ?

@ijingo @ug93tad @dloghin there are some configurations set in file /benchmark/ethereum/hosts and benchmark/ethereum/clients,
when I run
./run-bench.sh 4 32 4 8
I always encounter the following output and I am not sure if it is an error which will prohibit the generation of the final results
ssh: connect to host 10.0.0.100 port 22: Operation timed out done node 10.0.0.100

I am wondering if I need to modify the contents in /benchmark/ethereum/hosts and /benchmark/ethereum/clients to fit the conditions of my own computer? If every user needs to modify them, then according to what guidance should we modify?