P1-[4.0 bug hunting]-[Dashboard]-wrong memory usage info for TiFlash
Opened this issue · 17 comments
1. What did you do?
- deploy a cluster on Aliyun with tiup
- visit PD's dashboard -- cluster info -- hosts
2. What did you expect to see?
TiFlash's memory usage is correct
3. What did you see instead?
TiFlash's memory usage is incorrect
I have another cluster which deployed on Tencent Cloud, the TiFlash node on that also shows a memory usage at 96.2%
, but actually the node only has a memory usage at 65%
4. What version of TiDB are you using? (tidb-server -V
or run select tidb_version();
on TiDB)
root@node3:~# /tidb-deploy/tidb-4000/bin/tidb-server -V
Release Version: v4.0.0-rc.2
Edition: Community
Git Commit Hash: dfbb1ff8b0a8d55893bcea652748c40b9a9b91ad
Git Branch: heads/refs/tags/v4.0.0-rc.2
UTC Build Time: 2020-05-15 11:54:25
GoVersion: go1.13
Race Enabled: false
TiKV Min Version: v3.0.0-60965b006877ca7234adaced7890d7b029ed1306
Check Table Before Drop: false
/bug P1
/pick-up-challenge
/pick-up-challenge
/pick-up-challenge
/pick-up-challenge
/give-up-challenge
/pick-up-challenge
@you06 Hi, I want to work on this bug, but it seems like TiFlash
is not a open sourced repo?
I did some investigation, and according to this this design doc , the root cause must be in the server_info
provider in TiFlash
. Since TiFlash
is not open sourced, I guess I can't continue working on this issue.
@YKG I'm working on this bug. If there is anything you can make contribution to, I'll let you know.
/give-up-challenge
This issue has been fixed in TiFlash's latest commit.