Standby database fails to build

[hnwyllmm]

Environment

• Version: observer (OceanBase seekdb 1.2.0.0)
• Revision: 1-c1827b16ec3827425ba402d3012500551c4ecf68
• Build Branch: master
• Build Time: Mar 11 2026 16:21:46
• Build Flags: RelWithDebInfo|San...

Description

The standby database fails to build.

Steps to Reproduce

Not provided. This issue is consistently reproducible.

Creator: 15640940228-ctrl

[15640940228-ctrl]

We encountered the same issue in a new environment: a cascading standby scenario: A->B->C. A and B are in sync. A was deployed on the 11th, and B was also deployed on the 11th, but it was offline from the afternoon of the 11th to the afternoon of the 12th. After bringing B back online, we confirmed that A and B are synchronized (they can switch roles as primary/standby). When we deployed C, the standby build failed. The problematic environment (C) is: (

Environment for B: 6.12.234.221:/data/1/test/test.z1.obs0

[hnwyllmm]

This work item has been synced to the GitHub open-source repository: #207

[footka]

[Image]

The primary database is handling the disk read operations for the standby's I/O extremely slowly.

[footka]

[Image]

The primary database's disk read operations for handling I/O from the standby are exceptionally slow.

[footka]

Let's reproduce it one more time, and get the tsar monitoring up and running.

[footka]

Let's reproduce it one more time, and make sure the tsar monitoring is up and running.