dmemsys / fusee Goto Github PK
View Code? Open in Web Editor NEWThis is the implementation repository of our FAST'23 paper: FUSEE: A Fully Memory-Disaggregated Key-Value Store.
This is the implementation repository of our FAST'23 paper: FUSEE: A Fully Memory-Disaggregated Key-Value Store.
As I know, RACE has implemented SPLIT OPERATION to support extendible hashing, however, i find the SEARCH, UPDATE, INSERT, DELETE operation in FUSEE code except SPLIT.
I read it for a long time and still haven't found it, so hope to get your answer, thank you very much!
Hello authors, I am interested in FUSEE are trying to reproduce the ycsb bench on our servers.
However, I have a question regarding the split-workload.py. In particular, line 6 of the script mentions the existence of a ycsb-small/
directory containing small workload specifications. However, I was unable to locate this directory in the repository.
Would it be possible for you to provide me with this directory or help me locate it?
Thank you in advance for your time and assistance.
Hello, I am interested in FUSEE which is an admirable work! Recently I tried to run the source code, but met some erros and hope to get your help. My experimental environment was 3 servers with Mellanox-CX5 RNIC, using the ROCE v2 protocol; Two serve as Memory Nodes and one serves as Compute Node.
server 1's server_config is placed in /FUSEE/build/ directory, and discribled as follows:
{
"role": "SERVER",
"conn_type": "ENS7F1",
"server_id": 0,
"udp_port": 2333,
"memory_num": 2,
"memory_ips": [
"192.168.0.1"]
],
"ib_dev_id": 1,
"ib_port_id": 1,
"ib_gid_idx": 0,
"server_base_addr": "0x10000000",
"server_data_len": 15032385536,
"block_size": 67108864,
"subblock_size": 256,
"client_local_size": 1073741824,
"num_replication": 2,
"main_core_id": 0,
"poll_core_id": 1,
"bg_core_id": 2,
"gc_core_id": 3
}
server 2's server_config is placed in /FUSEE/build/ directory, and discribled as follows:
{
"role": "SERVER",
"conn_type": "ENS7F1",
"server_id":1,
"udp_port": 2333,
"memory_num": 2,
"memory_ips": [
"192.168.0.2"]
],
"ib_dev_id": 1,
"ib_port_id": 1,
"ib_gid_idx": 0,
"server_base_addr": "0x10000000",
"server_data_len": 15032385536,
"block_size": 67108864,
"subblock_size": 256,
"client_local_size": 1073741824,
"num_replication": 2,
"main_core_id": 0,
"poll_core_id": 1,
"bg_core_id": 2,
"gc_core_id": 3
}
server 3 as Compute node, and Client_config.json is placed in /FUSEE/build/ directory, and discribled as follows:
{
"role": "CLIENT",
"conn_type": "ENS7F1",
"server_id": 0,
"udp_port": 2333,
"memory_num": 2,
"memory_ips": [
"192.168.0.1",
"192.168.0.2"
],
"ib_dev_id": 1,
"ib_port_id": 1,
"ib_gid_idx": 0,
"server_base_addr": "0x10000000",
"server_data_len": 15032385536,
"block_size": 67108864,
"subblock_size": 256,
"client_local_size": 1073741824,
"num_replication": 2,
"num_idx_rep": 1,
"num_coroutines": 10,
"miss_rate_threash": 0.1,
"workload_run_time": 10,
"micro_workload_num": 10000,
"main_core_id": 0,
"poll_core_id": 1,
"bg_core_id": 2,
"gc_core_id": 3
}
Firstly, I run memory node as READme.md at server 1:
cd /FUSEE-master/build# numactl -N 0 -m 0 ./ycsb-test/ycsb_test_server 0
===== Starting Server 0 ====
kv_area_addr: 18000000, block_size: 4000000
my_sid_: 0, num_memory_:2
num_rep_blocks: 26, num_blovks: 26, limit: 90000000
press to exit
==== Ending Server 0 ====
Secondly, I run memory node as READme.md at server 2:
cd /FUSEE-master/build# numactl -N 0 -m 0 ./ycsb-test/ycsb_test_server 1
===== Starting Server 1 ====
kv_area_addr: 18000000, block_size: 4000000
my_sid_: 1, num_memory_:2
num_rep_blocks: 26, num_blovks: 26, limit: 90000000
press to exit
==== Ending Server 1 ====
Thirdly, I run compute node as READme.md at server 3:
cd /FUSEE-master/build/micro-test# numactl -N 0 -m 0 ./lantency_test_client ../client_config.json
main process running on core: 0
server_kv_area_addr: 28000000 26
num_rep_blocks: 26
Cannot have allocated 28000001
The error is " Cannot have allocated 28000001"
So, how to solve it? Could you tell me what each parameter in the servr_config.json and client_cofig.json is used for? Thank you very much!
Hi,你好,最近我尝试在两个节点上实现FUSEE,将nodeA作为CN,nodeB作为MN。在config.json文件里配置好IP等信息后开始编译运行。在nodeB上运行的是numactl -N 0 -m 0 ./ycsb-test/ycsb_test_server 0,以启动MN服务,在nodeA上运行的是numactl -N 0 -m 0 ./micro_test_multi_client ../client_config.json 8,以此测试吞吐。但是目前代码在client.cc的line132处遇到问题,具体原因是,while (!init_is_finished());当执行init_is_finished()函数里的ret = nm_->nm_rdma_read_from_sid(local_buf_, local_buf_mr_->lkey, sizeof(uint64_t), remote_global_meta_addr_, server_mr_info_map_[0]->rkey, 0);时,wc返回的status状态码是5,即Work Request Flushed Error,可以请您帮我一起解决这个问题吗?以及能否请教您为何会遇到这个错误。感谢。
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.