Will you be doing a long time experiment ? I used the same instructions to change a Pi3B (not +) with Jessie into an AP, worked for an hour or so and then it failed (repeatable). At the moment I am doing a long time experiment with Stretch, after 9 hours it has not yet failed. The symptoms are that several clients can connect initially but after some time hostapd reports that the stations disassociated and fails to reconnect, the AP is still active but does not accept new connections. The only way to recover is a reboot.
That's good to know. From what you say, it just works and is reliable. Do you have any idea how many devices can simultaneously connect to the access point? I'll be getting mine tomorrow and have a day to configure it before deploying it to an off-site location. The short time frame is a difficulty, so I appreciate your testing.
according to http://web.mit.edu/freebsd/head/contrib ... stapd.conf maximum 2007 stationsejolson wrote: ↑Wed Mar 28, 2018 5:58 pmThat's good to know. From what you say, it just works and is reliable. Do you have any idea how many devices can simultaneously connect to the access point? I'll be getting mine tomorrow and have a day to configure it before taking it with me.
Code: Select all
# Maximum number of stations allowed in station table. New stations will be # rejected after the station table is full. IEEE 802.11 has a limit of 2007 # different association IDs, so this number should not be larger than that. # (default: 2007) max_num_sta=255
There's a work-around posted on the forum and the RPF/RP(T)Ltd. folks are looking at it.ejolson wrote: ↑Wed Mar 28, 2018 7:54 pmThat doesn't sound so good. Anyway, I wasn't planning to bridge any devices, so hopefully it will just work as in your first test.
My Pi 3B+ arrived and I have set up an access point using the built-in WiFi on wlan0 with upstream through a USB WiFi dongle on wlan1. I didn't use bridging but NAT for a segmented network topology, thus avoiding the kernel panic reported above. Some scripting was needed to prevent dhcpcd from lauching wpa_supplicant on the access point. Also, since systemd refused to start the dhcp server, additional scripting was needed in rc.local for that and to set up the masquerade firewall. Any hints on how to troubleshoot why systemd does not start dhcp would be appreciated. Is there anything I need to do to turn off power saving so the access point doesn't go to sleep? I'll report back on reliability soon.
Switching the WiFi so the built-in was upstream definitely helped--the upstream signal was weak and USB dongle didn't have very good range. It also allowed me to track another issue with power saving modes. Everything is now running stably for a couple weeks!
The result shouldn't be unreliable, but I'll admit it was a bit unreliable.