# Flint 4 (4.11.0 beta 4): Mac randomly loses connection until wifi reconnect, router drops it silently after SA Query timeout **URL:** \ **Category:** Router\ **Created:** [October 17, 2026, 7:53pm UTC](https://forum.gl-inet.com/t/flint-4-4-11-0-beta-4-mac-randomly-loses-connection-until-wifi-reconnect-router-drops-it-silently-after-sa-query-timeout/72145 "2026-10-17T19:53:51Z")\ **Posts on this page:** 4\ **Page:** 1
**Author:** ![nemetharon](https://forum-assets.gl-inet.com/letter_avatar_proxy/v4/letter/n/7bcc69/32.png) [@nemetharon](https://forum.gl-inet.com/u/nemetharon)\ **Post date:** [October 17, 2026, 7:53pm UTC](https://forum.gl-inet.com/t/flint-4-4-11-0-beta-4-mac-randomly-loses-connection-until-wifi-reconnect-router-drops-it-silently-after-sa-query-timeout/72145/1 "2026-10-17T19:53:51Z")
Hi! I have another issue with my Flint 4 (main router) + Flint 3 (mesh node) setup, this time on firmware 4.11.0. My MacBook on wifi randomly loses network connection. macOS still shows it as connected to wifi, but nothing loads until I turn wifi off and on again. Other devices are fine during this. All 3 bands use the same SSID, but the security is different on each one: - 2.4 GHz: WPA2-PSK, PMF optional - 5 GHz: WPA2/WPA3 mixed, PMF optional - 6 GHz: WPA3 only, PMF required I went through the logs. The Mac moves between bands quite often. When it is connected on 5 GHz and tries to join 6 GHz, the router rejects it with status 30 and sends an SA Query to the old band. Most of the time it works out, but when the SA Query times out the router deletes the Mac from 5 GHz without sending a deauth. The Mac doesn't know it was dropped, so it keeps sending on 5 GHz and the router answers with "can't find pEntry". This goes on until I reconnect manually. Here is one case from today: 19:59:51 Mac connects on 5 GHz (rai0), gets DHCP, all fine 21:18:34 Mac tries 6 GHz (rax0) -\> auth resp status=30, SA Query sent 21:21:37 same 21:24:48 same 21:33:16 same, but this time: 21:33:18 PMF\_SAQueryTimeOut -\> mac\_entry\_delete, rai0: STA disassociated 21:35:47 BN1 Peer\_Vendor\_Spec\_Action: can't find pEntry 21:36:30 I reconnect manually, Mac joins 6 GHz again The same thing happened to my iPhone at 19:17 (status=30, then SA Query timeout, then dropped). The only devices that show this in the logs are Apple ones. Maybe that's because they switch bands more often than other devices. I think there are two problems here. First, the router deletes the client without telling it, so it ends up in a dead state instead of reconnecting by itself. Second, the security is different between bands on the same SSID, which probably makes the Mac do a full SAE handshake every time it changes band. I haven't changed anything yet. I attached the full log export. Can somebody please look into this? [logread-redacted.tar](https://forum.gl-inet.com/uploads/short-url/cAbDVw2unwcs1uGYvs7LNiN9PlN.tar) (769 KB) ---
**Author:** ![uvarvu](https://forum-assets.gl-inet.com/letter_avatar_proxy/v4/letter/u/fbc32d/32.png) [@uvarvu](https://forum.gl-inet.com/u/uvarvu)\ **Post date:** [October 18, 2026, 9:23am UTC](https://forum.gl-inet.com/t/flint-4-4-11-0-beta-4-mac-randomly-loses-connection-until-wifi-reconnect-router-drops-it-silently-after-sa-query-timeout/72145/2 "2026-10-18T09:23:10Z")
I have the same issue. I think I will commission the IoT network and boot everything that doesn't support WPA3 over to that. ---
**Author:** ![will.qiu](https://forum-assets.gl-inet.com/letter_avatar_proxy/v4/letter/w/b38774/32.png) [@will.qiu](https://forum.gl-inet.com/u/will.qiu)\ **Post date:** [October 22, 2026, 6:30am UTC](https://forum.gl-inet.com/t/flint-4-4-11-0-beta-4-mac-randomly-loses-connection-until-wifi-reconnect-router-drops-it-silently-after-sa-query-timeout/72145/3 "2026-10-22T06:30:01Z")
Hi, Thank you for the detailed report and logs. We will share this case with our R&D team for further investigation. To help us check the current wireless configuration, could you please send us a configuration [backup](https://forum.gl-inet.com/t/how-to-export-import-backup-configuration-files-via-luci/68190) of the router by private message? Please do not upload the backup publicly, as it may contain sensitive configuration information. Click the Technical Support avatar and then the **Message** button: ![image](https://forum-static.gl-inet.com/original/3X/1/0/10705d4a12191adc9fba880b69bfea7853d44efe.png) Please also let us know the exact MacBook and iPhone models and their macOS and iOS versions. If your account cannot start a private message, let us know and we can initiate one. ---
**Author:** ![chrismazanec](https://forum-assets.gl-inet.com/letter_avatar_proxy/v4/letter/c/eb8c5e/32.png) [@chrismazanec](https://forum.gl-inet.com/u/chrismazanec)\ **Post date:** [October 26, 2026, 8:07am UTC](https://forum.gl-inet.com/t/flint-4-4-11-0-beta-4-mac-randomly-loses-connection-until-wifi-reconnect-router-drops-it-silently-after-sa-query-timeout/72145/5 "2026-10-26T08:07:39Z")
**Same end state on Flint 4, different trigger (single-band WPA2 SSID, no roaming)** Adding a data point. GL-BE14000, firmware 4.9.2 and now 4.11.0 beta5 (build 1114), same behaviour on both. Client: Intel MacBook, macOS 13.7.8, Broadcom 802.11ac. It is on a 5 GHz-only SSID with WPA2-PSK (CCMP), PMF optional, so there is no band change or SAE involved. Symptom is identical to the OP: the Mac stays associated, no traffic passes, only Wi-Fi off/on recovers it. The AP logs nothing before the stall. After the client re-associates we see the same `PMF_SAQueryTimeOut`, `assoc ie is null` and `PN type mask is zero` lines, so in our case the SA Query timeout is a consequence, not the trigger. Frequency: 6–11 stalls per hour overnight, up to 13–18 per hour in the evening, for days. Also reproduced on 2.4 GHz and on a WPA3/MLO SSID. Things that did not change the rate (tested one at a time, 1 h each with a client-side watchdog counting stalls): WPA2-only instead of WPA2/WPA3 mixed, PMF off, `RROSupport=0`, DSCP-to-UP remap on/off, `ETxBfEnCond=0`, router reboot, 4.9.2 → 4.11.0 beta5. From AP-side counters on 4.9.2 during a stall: RX from the Mac continues, TX bytes towards the Mac stop and TX retry/drop counters grow (`Previous BA is not finished yet`, TxMsduDrop), BA window 64. So the AP appears to stop delivering downlink to that station while the station is still talking. Related but separate issue, already with support as Ticket-154243 and thread 72275: TCP marked DSCP AF21 (SSH SYN-ACK) never reaches the client while everything else works. Full logread, client logs and the test matrix are available; happy to test builds or collect specific debug output.