To improve cluster stability, upgrade the ACK Node Problem Detector (ack-node-problem-detector) add-on to version 1.2.8 or later.
Background information
Versions of ack-node-problem-detector earlier than v1.2.8 enable the file system handle check (check_fd) by default. When the check_fd process reads large amounts of kernel data, it can trigger a kernel bug in certain kernel versions. This may cause zombie processes to accumulate or even lead to system downtime.
Upgrade the add-on
-
To prevent issues that are caused by earlier versions, upgrade the ack-node-problem-detector add-on to version 1.2.8 or later. This version disables the file system handle check (check_fd) by default and prevents the process from reading large amounts of kernel data. For more information, see ack-node-problem-detector.
-
Fix the kernel bug. For more information, see Page fault exception causes system downtime on ECS instances running Alibaba Cloud Linux 2 .
Log on to the ACK console. In the left navigation pane, click Clusters.
On the Clusters page, click the name of your cluster. In the left navigation pane, click Add-ons.
-
On the Add-ons page, go to the Logs and Monitoring tab. Click ack-node-problem-detector, and then click Upgrade in the lower-right corner.
NoteIf the Upgrade button is not displayed on the ack-node-problem-detector panel, the add-on is already the latest version and an upgrade is not required.
-
In the confirmation dialog box, click OK.