etcd perforamance limit #9112
Labels
Comments
|
We successfully simulated 5Mi pods case. It seems to me that you did something wrong. Please share the scripts(etcd setup, put test) to reproduce your test, and please share the stack trace for the crash. |
|
@xiang90 Thanks for your reply 2.Here's some info maybe useful |
|
etcd start up command: |
well... it runs out of memory. so you need to provide more memory i guess. please share your scripts to run the test if you want anyone to look into it more. |
|
@xiang90 hello,we want to confirm the count of pod. What 5Mi pods mean |
|
it means 5,000,000 kvs. |
|
insert_pod.sh: insert_audit.sh: get_json.sh: |
|
Your script tests etcd2 backend which supports 200MB at most, not etcd3. We have no plan to scale etcd2 backend, and k8s uses etcd3 by default in its recent versions. Closing this issue. |
|
@xiang90 how should I use etd3? |
|
yes |
|
@xiang90 hello~ please give me some help or info. |
|
@WanLinghao Can you please do some simple calculation on your own? And try to figure out if the db size matches with what you put into it? We are not going to solve all your problem for you. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
We are doing a etcd performance limit test recently.The test goes like this:
Insert some data into an etcd cluster, then test its performance, such as read time consume.
At first, we want to insert 500,0000 data into the cluster.But the cluster collapsed at 50,0000 data during test process.
The test data is all about kubernetes info, all data follows the pattern below
We want to know where's the limit of etcd cluster? Did we hit the limit of etcd cluster?
If not, anything wrong we did?
Here's some information maybe useful.
etcd Version: 3.1.3
3-node cluster, each node has 4GB memory.
The text was updated successfully, but these errors were encountered: