OP25 Firmware for Flint 2 & Beryl AX is out

It’s solved, it was an Openwrt Adblock list (1hosts Var compilation) that was blocking the default ddns server address of the Flint 2. Thank you all for the tips :+1:

3 Likes

Hi,

Sorry for the delayed reply, and thank you for the detailed report.

Could you please check whether the same issue also occurs on the stock firmware?

Based on the related error messages, this appears to be an issue with the upstream Vanilla OpenWrt open-source Wi-Fi driver, so we may need to wait for the upstream project to fix it before we can merge the corresponding changes:

If you confirm that the issue also occurs on the stock/closed-source firmware, please export the logs after the issue occurs and send them to us via private message.

We will investigate further.

1 Like

Hi,

Could you please provide the following details so we can investigate the issue further?

  1. When the issue occurs, is the 2.4GHz Wi-Fi channel set to Auto or a fixed channel?
  2. Does the issue affect all wireless devices, or only a specific device?
  3. If possible, when the issue occurs again, please export the logs and send them to us via private message for further investigation.
How to export logs:

How to send private messages:

1 Like

I've never used stock firmware, but this issue didn't occur during the past year on op24 firmware.

Thanks for your prompt response.

Since this occurred only once after a long uptime, could you first confirm whether it can now be reproduced consistently and whether there is a repeatable trigger or an approximate time to failure?

If the issue can be reproduced reliably, we recommend comparing the behavior with both the regular GL.iNet firmware and an official Vanilla OpenWrt build for the GL-MT6000.

Before doing so, you can back up your configuration first so that you can quickly restore it when switching back. Please do not keep or restore settings when switching between these different firmware branches.

If the issue occurs on Vanilla OpenWrt but not on the regular GL.iNet firmware, it would indicate an upstream OpenWrt driver issue. In that case, the fix would need to come from the upstream project before we could integrate it.

I just want to make sure you understand, that I have not used either non-beta GL.iNet, nor vanilla OpenWRT. I used op24 from GL.iNet for about a year w/o encountering this issue and now I’m using op25 beta3 (4.9.1 op25 b3). If the issue does occur again, it may take 19 days like it did the first time. So far my uptime is 14days. So saying this issue can be reproduced reliably is essentially hoping that after 19 days of uptime, it happens again.

The op24 always been a "beta" firmware its never been a "stable" release

Didn’t I say:

Roger that, my bad :call_me_hand:

Thank you for clarifying. We understand that you have only used GL.iNet OP24 and OP25 beta3, and that the issue has occurred only once after approximately 19 days.

Our previous reply was describing the comparison steps that would be needed to identify where the issue originates. We did not mean that you had already tested the regular GL.iNet firmware or Vanilla OpenWrt, or that the issue was currently reproducible.

If the issue becomes reliably reproducible, comparing it with both firmware branches would help distinguish an OP25 integration issue from an upstream OpenWrt driver issue. Without those comparison results, it may be difficult to determine which component is responsible, especially given the long and currently uncertain time to failure.

My understanding of op24 and op25 was that you take vanilla OpenWRT, add GL.iNet interface with a few addons and you also change the Ethernet/WiFi drivers.

Given that I didn’t have this issue with op24 for about a year, I would assume this is likely the result of OpenWRT v25 and/or changes you pushed to op25 vs op24.

Given the bug report with similar issues on vanilla v25, I would assume this is an OpenWRT v25 issue, which I don’t think they fixed yet.

In case anyone has an issue, bark parental controls is currently broken in op25. I’ve had to back off to op24 to get it fixed. I am told the current apk used is not compatible, and they are working with bark to get a new one.

I also reverted to op24. On op25 i receive this error GL-MT6000 WAN random crash - only recoverable by rebooting · Issue #17624 · openwrt/openwrt · GitHub, or similar but also NETDEV WATCHDOG: eth0 (mtk_soc_eth): transmit queue 5 timed out . Usually it happens when i use more than 50Mbps on WAN. But it also worked for 2 weeks fine but then started failing again and only reboot makes the interfaces working again, but instantly crashes again.

Hi,

This appears to be related to the following upstream OpenWrt issue:

We sent you a debug firmware build via private message. Could you please install it and check whether the issue still occurs?

If it does, please try disabling GRO completely to see whether that helps:

ethtool -K eth0 gro off
ethtool -K eth1 gro off

You can re-enable GRO afterward with:

ethtool -K eth0 gro on
ethtool -K eth1 gro on

or reboot the router.

Just had a complete freeze. Got this from the logs:
[Sep 8, 2026, 10:33:08 PM EDT] kern.err: [ 1.765542] u-boot-env-layout 11230000.mmc:card@0:block:partitions:block-partition-env:nvmem-layout: Invalid calculated CRC32: 0xc09e9b4c (expected: 0x00000000)
[Sep 8, 2026, 10:33:08 PM EDT] kern.err: [ 1.779915] u-boot-env-layout 11230000.mmc:card@0:block:partitions:block-partition-env:nvmem-layout: probe with driver u-boot-env-layout failed with error -22
[Sep 8, 2026, 10:33:08 PM EDT] user.err: [ 7.266730] block: unable to load configuration (fstab: Entry not found)
[Sep 8, 2026, 10:33:08 PM EDT] user.err: [ 7.279980] block: unable to load configuration (fstab: Entry not found)
[Sep 8, 2026, 10:33:08 PM EDT] user.err: [ 7.293105] block: unable to load configuration (fstab: Entry not found)
[Sep 8, 2026, 10:33:08 PM EDT] user.err: [ 7.299846] block: no usable configuration
[Sep 8, 2026, 10:33:08 PM EDT] user.err: [ 7.309356] block: unable to load configuration (fstab: Entry not found)
[Sep 8, 2026, 10:33:08 PM EDT] user.err: [ 7.316060] block: no usable configuration
[Sep 8, 2026, 10:33:08 PM EDT] user.err: [ 7.532895] block: unable to load configuration (fstab: Entry not found)
[Sep 8, 2026, 10:33:08 PM EDT] user.err: [ 7.539608] block: no usable configuration
[Sep 8, 2026, 10:33:13 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Network unreachable
[Sep 8, 2026, 10:33:13 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Network unreachable
[Sep 8, 2026, 10:33:13 PM EDT] daemon.err: cellular_manage: (slot_failover.c:963) No 'bcpu_p_v' found in slot_feature config
[Sep 8, 2026, 10:33:13 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Network unreachable
[Sep 8, 2026, 10:33:13 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:13 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:13 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:14 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:14 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:14 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:14 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:14 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:14 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:14 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: nmbd[5450]: [2026/09/08 22:33:15.180293, 0] ../../source3/nmbd/nmbd.c:901(main)
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: nmbd[5450]: nmbd version 4.18.8 started.
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: nmbd[5450]: Copyright Andrew Tridgell and the Samba Team 1992-2023
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: smbd[5448]: [2026/09/08 22:33:15.181541, 0] ../../source3/smbd/server.c:1746(main)
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: smbd[5448]: smbd version 4.18.8 started.
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: smbd[5448]: Copyright Andrew Tridgell and the Samba Team 1992-2023
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: hostapd: rmdir[ctrl_interface=/var/run/hostapd]: Permission denied
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: hostapd: hostapd_free_hapd_data: Interface wlan0-1 wasn't started
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: hostapd: rmdir[ctrl_interface=/var/run/hostapd]: Permission denied
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: hostapd: hostapd_free_hapd_data: Interface wlan0 wasn't started
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:15 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:16 PM EDT] daemon.err: procd: Got unexpected signal 1
[Sep 8, 2026, 10:33:16 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:16 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:16 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:16 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:16 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:16 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:17 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:17 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call vpn-client.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:17 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:20 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call vpn-client.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:20 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:21 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:23 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call vpn-client.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:23 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call cable.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:25 PM EDT] daemon.err: eco[2267]: (gl-ngx-session:306) call vpn-client.get_status fail: connect "127.0.0.1 8081" fail: Connection refused
[Sep 8, 2026, 10:33:33 PM EDT] daemon.err: nmbd[5450]: [2026/09/08 22:33:33.463281, 0] ../../source3/nmbd/nmbd.c:59(terminate)
[Sep 8, 2026, 10:33:33 PM EDT] daemon.err: nmbd[5450]: Got SIGTERM: going down...

Hi,

Thank you for your feedback.

However, the current logs only show the following:

  1. The fstab: Entry not found messages mean that OpenWrt could not find a usable fstab configuration during startup. Could you confirm whether you made any changes to fstab, such as configuring extroot or mounting an external drive as /overlay?

  2. The subsequent messages indicate that our internal eco component could not communicate with Nginx. Could you check whether the relevant processes are running normally?

    ps | grep -E '[n]ginx|[e]co'
    

If possible, please enable persistent logging and, after the issue occurs again, export the /root/system.log and router logs send them to us.

How to export logs:

How to send private messages:

Hello,

Before the original crash, I had not configured extroot or made any custom fstab changes. The router was using its normal internal overlay.

After receiving your recommendation, I attached a dedicated USB SSD and configured persistent logging there instead of /root, to avoid continuous writes to the router’s internal flash. The logger now writes to an ext4 filesystem with a 256 MiB rotation size, while remote syslog to my Synology remains enabled.

I performed another controlled reboot and collected the following observations:

1. The same U-Boot environment CRC error appears on every boot:

u-boot-env-layout ... Invalid calculated CRC32: 0xc09e9b4c

u-boot-env-layout ... probe with driver u-boot-env-layout failed with error -22

This exact MT6000/25.12 issue has already been reported and marked as a confirmed upstream OpenWrt bug:

Since it is consistently reproducible on otherwise successful boots, it appears separate from the unexpected router crash.

2. At the time of the original crash, /etc/config/fstab existed and contained the stock global settings plus the two default disabled /overlay and /rom mount entries. I had not modified it, configured extroot, or configured an external drive.

I verified this against a router backup made before the incident. The fstab: Entry not found messages therefore occurred despite the stock fstab being present.

After the incident, I added one UUID-based ext4 mount solely to store the persistent diagnostic log requested by support. This change was not present when the crash occurred.

On the latest controlled reboot, the same fstab: Entry not found messages appeared around 7.3–7.7 seconds. Nevertheless, the configured external ext4 filesystem mounted successfully at approximately 11.7 seconds. This suggests an early boot or overlay timing issue rather than an absent or unusable final fstab configuration.

3. The eco errors are transient during startup. On the controlled reboot, they occurred while connections to 127.0.0.1:8081 were refused. Once startup completed:

  • The Nginx master and worker processes were running.

  • Nginx was listening on IPv4 and IPv6 port 8081.

  • gl-ngx-session, gl-repeater, and gl-dpi eco processes were running.

  • The connection-refused messages stopped.

4. I reviewed the remote Synology syslog surrounding the original crash. It showed routine activity through approximately 22:31, DHCP traffic around 22:35–22:36, WAN offline at 22:45:25, and DNS-over-TLS timeouts at 22:45:36 and 22:45:41. There was then a logging gap followed by device reconnections around 22:52. I found no kernel panic, OOM kill, watchdog report, call trace, or similar crash signature. The router’s ramoops/pstore area was also empty after the incident.

Persistent USB logging is now enabled so that another occurrence should leave more local evidence. Please advise whether there are any additional debug facilities, watchdog diagnostics, or GL.iNet-specific crash dumps you would like enabled to identify the cause of the unexpected crash.

Hi,

Thank you for the detailed update and for sharing your analysis.

Since the same fstab , U-Boot environment, and transient eco messages also appear during normal boots, they are unlikely to explain the freeze. The previously provided excerpt appears to show the normal startup sequence after the incident rather than the events that caused it.

The system log includes both system and kernel messages by default, so your current persistent USB logging together with remote syslog should normally provide enough information for further analysis. No additional debugging needs to be enabled at this stage.

If the issue occurs again, please send us the logs covering the period immediately before and after the failure. We will review them with R&D team.

OP25 4.9.1 Beta 3 came out on 2026-07-22.

Looking at Snapshot for Flint 2 - 2026-09-15 - there have been a LOT of bug fixes since 07-22:

2026-09-14: Added support for unquoted APNs in CGCONTRDP responses.
2026-09-11: Added source interface validation before creating IoT Wi-Fi networks.
2026-09-11: Added support for domains beginning with digits in VPN subscriptions.
2026-09-10: Added support for PSK SHA-256 repeater connections.
2026-09-10: Fixed web language update failures on screen devices.
2026-09-10: Fixed UBI errors caused by power loss after moving files.
2026-09-10: Fixed guest and IoT interface errors after upgrades with preserved configurations.
2026-09-10: Fixed 6 GHz repeater scan timeouts in dense environments.
2026-09-09: Fixed AP isolation after holding the reset button for four seconds to reset the network.
2026-09-09: Fixed repeater connections so band locking does not interrupt an existing connection.
2026-09-09: Fixed five-second scan delays when creating interfaces with nl80211.
2026-09-09: Fixed repeater bandwidth mismatches between 2.4 GHz AP and STA interfaces.
2026-09-09: Fixed concurrent scans entering an invalid state that prevented subsequent scans.
2026-09-09: Fixed scan failures when the STA interface was down.
2026-09-09: Fixed scan failures caused by creating the sta0 interface without a MAC address.
2026-09-09: Fixed the 415 compilation error.
2026-09-08: Synchronized screen language updates with the web interface.
2026-09-08: Mapped anonymous guest sessions to nobody.
2026-09-04: Prevented duplicate classification of askpass configurations.
2026-09-03: Preserved secondary IPv4 addresses.
2026-09-03: Fixed guest Wi-Fi connection failures on X2000.
2026-09-03: Fixed IoT Wi-Fi connection failures on X2000.
2026-09-02: Set disk_number to 0 when no disk is present.
2026-09-02: Prevented MT3000 from using the DE SKU table.
2026-09-01: Added backhaul profile parameter validation.
2026-09-01: Fixed slow or empty NAS folders when accessed remotely through Tailscale with the iOS WebDAV client.
2026-08-31: Added AzireVPN credential validation before saving.
2026-08-31: Fixed automatic repeater MAC address changes after disconnection.

When - if needed - will these bug fixes be rolled into OP25 firmware?

I also see the latest Beta firmware 4.11 adds a lot of new features - OP25 missing out at present.

Hi,

Thank you for your concern.

First, we would like to clarify that snapshot changelogs cover development across multiple models and components, so not every listed change necessarily applies to Flint 2 or Beryl AX.

We are currently focused on releasing the stable v4.11 firmware for Flint 2 and Beryl AX. Only after that may we begin working on v4.11.x-op25.