Time to create an overlay netowrk significantly increased as the number of agents increases.
Nobody has claimed this yet.
- Dominant language
- Go
- Stars
- 2.2k
- Forks
- 875
- PR merge metrics
- No merged PRs in 30d
Description
Hi, I compare the time creating an overlay network in a libnetwork cluster with different number of dnet.
In a 50-dnets setup, it take ~ 1s, which looks reasonable.
# time ./bin/dnet -H tcp://127.0.0.1:2385 network create -d overlay multihost_50
45ed34c232aa40194a15a0da64432fc25b403780b6f1ea9bfee8ec5a650a174e
real 0m1.540s
user 0m0.012s
sys 0m0.008s
However, in a 100-dent cluster it takes about 10s
# time ./bin/dnet -H tcp://127.0.0.1:2385 network create -d overlay multihost_100
3e28785cd3d5fd1b8cfd8574b8511053e356216a101652a7e0fe641aff5cc847
real 0m10.098s
user 0m0.004s
sys 0m0.012s
The idle CPU is 90%, so the system is not saturated.
In my understanding, the network resource has global scope.
So besides watching for the KV, does the overlay driver do something else with the network KV which slows down this operation? Thanks.
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Reproduce the reported ./bin/dnet -H tcp://127.0.0.1:2385 network create -d overlay timings with 50- and 100-dnet clusters. Start by tracing what the overlay driver does with the network KV beyond watching it; done means identifying the source of the scaling slowdown and verifying the explanation against both measurements.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- go
- Domain
- networking
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100