We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
pinger 默认每5秒调用一次 ovn-sbctl 查询,当集群规模越来越大,对 ovn-sb-leader 的造成非常大的压力。以下对比 ovn-sbctl 命令和 ovsdb-client 命令的在大量请求情况下 ovn-sb-leader 的压力: test-ovn-sbctl.sh
test-ovn-sbctl.sh
#!/bin/bash while true; do ovn-sbctl --db=tcp:$OVN_SB_SERVICE_HOST:6642 --format=csv --no-heading --data=bare --columns=_uuid --timeout=10 find chassis hostname=192.168.212.16 > /dev/null & ovn-sbctl --db=tcp:$OVN_SB_SERVICE_HOST:6642 --format=csv --no-heading --data=bare --columns=logical_port --timeout=10 find port_binding chassis=22f903d0-35a2-4f7c-ad6f-1bc66c4c08c8 > /dev/null& sleep 0.05 done
脚本模拟 100 个节点的 pinger 请求,当我并发启动5个 test-ovn-sbctl.sh 脚本时(等价于模拟500个节点),cpu 已经占用到 60% 以上,在负载较大的k8s表现会更糟糕。
当修改使用ovsdb-client时,这种情况得到明显改善。 test-ovsdb-client.sh
test-ovsdb-client.sh
#!/bin/bash while true; do ovsdb-client --timeout=10 query tcp:$OVN_SB_SERVICE_HOST:6642 '["OVN_Southbound", {"op": "select","table": "Chassis","where": [["hostname", "==", "192.168.212.16"]],"columns": ["_uuid"]}]' > /dev/null & ovsdb-client --timeout=10 query tcp:$OVN_SB_SERVICE_HOST:6642 '["OVN_Southbound",{"op": "select","table": "Port_Binding","where": [["chassis", "==", ["uuid","22f903d0-35a2-4f7c-ad6f-1bc66c4c08c8"]]],"columns": ["logical_port"]}]' > /dev/null& sleep 0.05 done
同样的,我们并发启用5个test-ovsdb-client.sh脚本,cpu 没有压力的表现
改善大规模集群下ovnsb-leader的压力问题
No response
The text was updated successfully, but these errors were encountered:
Successfully merging a pull request may close this issue.
Description
pinger 默认每5秒调用一次 ovn-sbctl 查询,当集群规模越来越大,对 ovn-sb-leader 的造成非常大的压力。以下对比 ovn-sbctl 命令和 ovsdb-client 命令的在大量请求情况下 ovn-sb-leader 的压力:
test-ovn-sbctl.sh
脚本模拟 100 个节点的 pinger 请求,当我并发启动5个
test-ovn-sbctl.sh
脚本时(等价于模拟500个节点),cpu 已经占用到 60% 以上,在负载较大的k8s表现会更糟糕。当修改使用ovsdb-client时,这种情况得到明显改善。
test-ovsdb-client.sh
同样的,我们并发启用5个
test-ovsdb-client.sh
脚本,cpu 没有压力的表现Who will benefit from this feature?
改善大规模集群下ovnsb-leader的压力问题
Anything else?
No response
The text was updated successfully, but these errors were encountered: