Mirror entered trouble state

Question

Question

Igor Pak · Dec 12, 2024

Dear community, my company use intersystems iris, a old version of 2019 and centos 7.9

At the moment, we are faced with a rather strange luck of iris.

Every day during working hours, we receive a error that iris primary cannot transfer the journal to the backup node.

We have availability monitoring, these errors do not affect availability in any way, but I still wanted to analyze this error.

I understand that the severity of this error is only 1, but Can someone share a diagnostic method this error?

12/12/24-15:48:43:705 (20263) 1 MirrorServer: Unlinking with journal daemon at 0x09db5294, returning to catchup state.

12/12/24-15:48:43:705 (15734) 1 MirrorServer: Mirror entered trouble state (backupjrnend = 138522624)

12/12/24-15:48:43:705 (15734) 1 MirrorServer: Mirror id #0 set trouble state. Failed to send journal data through 138555856 for file #51647 (cur = 51647/138522624)within timeout period. Disconnected from journal system

12/12/24-15:48:43:762 (20288) 0 MirrorServer: Ping ack received from unlinked backup...(repeated 1 times)

12/12/24-15:48:43:762 (20288) 0 MirrorServer: Mirror trouble state cleared. Switched to Agent Controlled failover.

12/12/24-15:48:43:763 (20288) 0 MirrorServer: Ping ack received from unlinked backup

12/12/24-15:48:43:805 (20263) 0 MirrorServer: Client up to date, linking with journal daemon @ 0x09db5294

12/12/24-15:48:43:806 (20263) 0 MirrorServer: Sending Prepare to Switch Modes

12/12/24-15:48:43:807 (20263) 0 MirrorServer: Switched from Agent Controlled to Arbiter Controlled failover

Product version: IRIS 2019.1

$ZV: 2019.1.1 (Build 612U)

Discussion (1)2

Log in or sign up to continue

score 1 · Answer 1 · 2025-01-02T08:29:55-05:00

Igor, from those messages, I would surmise that you had some sort of brief network disconnect which ultimately resolved itself. The primary mirror member, of course, will continue to support your workload so availability would not be affected.

This doesn't look like a problem with IRIS, but rather a response to the conditions on the network.

You might also reach out to the WRC to determine if there is something about your specific version that is causing these brief disconnects to be reported.