Thanks to everyone who responded, and yes, I managed to make this much more difficult than it should have been!
As a newbie to aruba, I've been practicing Just In Time Learning (JITL). Since I have a controller-based cluster I know little beyond the basic concepts. I don't have a proper isolated lab setup and so am just running this on my live production environment. I managed to get a couple of totally wrong ideas along the way, and once I realized that and had my facepalm-with-an-octopus, it's really pretty easy.
I learned a lot through the process, and maybe this will help others learn :-)
Here's my eventual process:
-- My husband built me a console cable, without power. (The real arubanetworks cable is out of stock everywhere.)) If you look at the console port with the AP right side up, the bottom pin in the ground, and the next two are transmit and receive. Top is power, don't hook that up and control power via the ethernet port. Hook ground to ground on the adaptor, and TXD<-->RXD and RXD<-->TXD. Plug the cable into the USB port of the computer. I use windows and PuTTY, and found the cable hooked into COM3.
-- Place the AP on its edge with the reset button on top. Use a paperclip to hold the reset button down, and click the ethernet cable into eth0 on the back of the AP, using a POE port where you can see your controller. The console will start rolling by on by, and should see "**** Configuration Reset Requested by User ****" and it clearing the flash, etc. This takes a few seconds and you can let go of the paperclip after you see "Hit <Enter> to stop autoboot:"
-- The AP then boots into instant mode, and it takes awhile. There is steady chatter in the console window as it goes, a lot of it interesting. The console will announce that the AP is up and give a "User:" prompt but it's not really up yet. You will see "i am master now" and the ethernet ports come up, and when you see "Sync Configuration Completed!" it's really up. Username and password is admin/admin.
-- Now that the master is up, add the other IAPs to the switch to form an instant cluster. (I only have 8 POE ports mapped to the right VLAN on this switch so I did this in batches.) Add each one the way you did the first one, laying it on it's edge and holding your paperclip in the reset button and plugging in the ethernet cable. Once you are sure that they are all up because the flashing green lights go to solid green, return to your computer and the console that's hooked into the master.
-- Once you are logged in, your prompt is the MAC address of the AP. At this point, I ping a controller's IP address to make sure I can see it. Then issue an 'enable' command followed by 'convert-aos-ap <IP address of your controller>'
20:4c:03:10:78:44# ping 10.10.31.208Press 'q' to abort.PING 10.10.31.208 (10.10.31.208): 56 data bytes64 bytes from 10.10.31.208: icmp_seq=0 ttl=64 time=2.9 ms64 bytes from 10.10.31.208: icmp_seq=1 ttl=64 time=0.4 ms64 bytes from 10.10.31.208: icmp_seq=2 ttl=64 time=0.4 ms64 bytes from 10.10.31.208: icmp_seq=3 ttl=64 time=0.4 ms64 bytes from 10.10.31.208: icmp_seq=4 ttl=64 time=0.4 ms--- 10.10.31.208 ping statistics ---5 packets transmitted, 5 packets received, 0% packet lossround-trip min/avg/max = 0.4/0.9/2.9 ms20:4c:03:10:78:44# enable20:4c:03:10:78:44# convert-aos-ap CAP 10.10.31.208Converting to Controller based AP. Will automatically reboot when done.This converts ALL of the IAPs in the instant cluster. The conversion should take about 2 minutes, and then the APs all reboot as campus APs with the ArubaOS on the controllers.
The wrong idea that I got in my head was that I needed to boot up the AP when isolated from the network, and then plug in after the AP was totally booted. (I did this using a POE injector and not connecting the network to the injector until later.) The way this failed is that the AP would just sit at the "Will automatically reboot when done." for hours, days, or forever. What made it hard was that the two minutes that it sits there with no activity is a very long time, and I wasn't waiting long enough before deciding that something was wrong. Because of the way networks work, adding the network to the AP after bootup would put it in an essentially random mixture of broken and working -- very confusing!
What also flummoxed me was that I got a couple of APs to convert successfully on a Friday afternoon, came down with covid over that weekend, and was out of the office for 10 days. Oh, and when I rebooted my windows computer to install the driver for the console cable, my Microsoft Edge history got deleted. (I hate windows!!!) It turns out that lots of things that I thought I knew were actually a matter of being able to locate the web page I was on, and without a browser history I can't find a bunch of stuff. So "don't use Edge for a web browser" is another piece of advice I'm handing out!
Upshot is that I learned a lot, and it was reasonably straightforward once I understood what was going on...
Original Message:
Sent: Jun 06, 2022 06:22 PM
From: Colin Joseph
Subject: Converting instant APs to campus APs -- the struggle is real
To tell the truth, you can boot all of the instant APs and wait until they all auto join a cluster and you can convert them all at the same time through the maintenance tab. Have you tried that?
------------------------------
Any opinions expressed here are solely my own and not necessarily that of Hewlett Packard Enterprise or Aruba Networks.
HPE Design and Deploy Guides: https://community.arubanetworks.com/support/migrated-knowledge-base?attachments=&communitykey=dcc83c62-1a3a-4dd8-94dc-92968ea6fff1&pageindex=0&pagesize=12&search=&sort=most_recent&viewtype=card
Original Message:
Sent: Jun 05, 2022 05:52 PM
From: Cathy Fasano
Subject: Converting instant APs to campus APs -- the struggle is real
I have 36 IAP 205Hs that I need to convert to campus APs.
I have successfully converted 11 of them.
I have a console cable (home built, as these seem to be out of stock everywhere.) The console has transmit, receive and ground hooked up, and power comes from a POE injector.
I start by doing a factory reset on the IAP (I don't have passwords.)
-- Plug in power to the POE injector but NOT network, plug in eth0: to the injector
-- Hit enter to interrupt the boot, then the series of commands to do the reset:
Hit <Enter> to stop autoboot: 0
apboot> purgeenv
preserving os_partition (1)
Erasing flash...Writing to flash..................done
Erasing flash...Writing to flash..................done
apboot> clear os
Erasing flash sector @ 0x1e100000....done
Erased 1 sectors
apboot> save
Saving Environment to Flash...
Erasing flash...Writing to flash..................done
apboot> factory_reset
Clearing state... Checking OS image and flags
Invalid image format version: 0xffffffff
Continuing with OS clear
Erasing flash sector @ 0x1e100000....done
Erased 1 sectors
Erasing flash sector @ 0x1ff80000....done
Erased 1 sectors
Erasing flash sector @ 0x1ff90000....done
Erased 1 sectors
Erasing flash sector @ 0x1ffb0000....done
Erased 1 sectors
Erasing flash sector @ 0x1ffc0000....done
Erased 1 sectors
Erasing flash sector @ 0x1ffd0000....done
Erased 1 sectors
done
Purging environment... preserving os_partition (1)
Erasing flash...Writing to flash..................done
Erasing flash...Writing to flash..................done
done
apboot> ▒
-- at this point I unplug/replug into eth0: to reboot the IAP. Lots and lots of console chatter, including an ethernet timeout (expected since the POE injector is not plugged into the network), and after 3 minutes I get a User prompt, but it's not ready for about another 3 minutes, then I can log in -- username/password is admin/admin. It is clearly coming up as an instant, starting with
[ 0.000000] Alcatel-Lucent
[ 0.000000] AOS-W Version 6.4.4.4-4.2.3.0 (build 54225 / label #54225)
[ 0.000000] Built by p4build@rhodes on 2016-03-17 at 03:53:14 PDT (gcc versio n 4.7.2)
[ 0.000000]
[ 0.000000] Flash variant: variant0
and finally
<<<<< Welcome to the Access Point >>>>>
Completed SW FIPS KAT test
User: [ 129.580000] Picked up default IP a9fe6c94, rand 6c94!
[ 135.716000] Default IP is ready!
[ 139.386000] ip_time_handler: Got ip and packets on bond0 Started master election 1-0, rand 22
[ 161.408000] i am master now
[ 161.410000] (08:02:38) !!! Init ---> Master
[ 161.412000] asap_send_elected_master: sent successfully
[ 169.552000] ethernet_device_event: dev eth1 others: 13
[ 169.556000] ADDRCONF(NETDEV_UP): eth1: link is not ready
[ 169.558000] ethernet_device_event: dev eth1 is up
[ 170.214000] ethernet_device_event: dev eth2 others: 13
[ 170.216000] ADDRCONF(NETDEV_UP): eth2: link is not ready
[ 170.218000] ethernet_device_event: dev eth2 is up
[ 170.884000] ethernet_device_event: dev eth3 others: 13
[ 170.886000] ADDRCONF(NETDEV_UP): eth3: link is not ready
[ 170.888000] ethernet_device_event: dev eth3 is up
Sync Configuration Completed!
At this point I plug in the network to the POE injector and console says:
[ 209.228000] eth0: link up (1000FD)
[ 209.230000] ethernet_device_event: dev eth0 is changed, carrier: 1
[ 209.610000] bonding: bond0: link status definitely up for interface eth0.
[ 209.612000] bonding: bond0: making interface eth0 the new active one.
[ 209.614000] bonding: bond0: first active interface up!
[ 209.616000] ADDRCONF(NETDEV_CHANGE): bond0: link becomes ready
[ 213.112000] asap_send_elected_master: sent successfully
I log in, and ping my controller. Then issue the 'enable' command, followed by the 'convert-aos-ap CAP <IP-of-controller>' command.
20:4c:03:10:73:34# ping 10.10.31.208
Press 'q' to abort.
PING 10.10.31.208 (10.10.31.208): 56 data bytes
64 bytes from 10.10.31.208: icmp_seq=0 ttl=64 time=2.3 ms
64 bytes from 10.10.31.208: icmp_seq=1 ttl=64 time=0.4 ms
64 bytes from 10.10.31.208: icmp_seq=2 ttl=64 time=0.3 ms
64 bytes from 10.10.31.208: icmp_seq=3 ttl=64 time=0.4 ms
64 bytes from 10.10.31.208: icmp_seq=4 ttl=64 time=0.4 ms
--- 10.10.31.208 ping statistics ---
5 packets transmitted, 5 packets received, 0% packet loss
round-trip min/avg/max = 0.3/0.7/2.3 ms
20:4c:03:10:73:34# enable
20:4c:03:10:73:34# convert-aos-ap CAP 10.10.31.208
Converting to Controller based AP. Will automatically reboot when done.
Here's the struggle... If this is running smoothly, about 2 minutes later that console will come back to life and the green light will flash, and it will reboot as a campus AP. But then sometimes it doesn't take just two minutes. I've had it take 4 hours, and then suddenly notice that the AP has booted. I've left it overnight, and come back and it had rebooted overnight. On Friday I got 4 or 5 APs converted reasonably quickly, and then by the end of the day it just hung at "Will automatically reboot when done." When I came in this afternoon 2 days later it was still stuck there.
If I reboot the AP it comes up in instant mode with the factory reset OS.
When I log into the AP it suggests the commands 'show tech-support' and 'show tech-support supplemental' as commands. I can run those commands while the AP is sitting at "Will automatically reboot when done." and it gives a tsunami of information.
Can anyone explain how to make this work more reliably? Is there anything I can look for? Is the controller maybe not answering? Any way the kick that and get its attention?
I'm really stumped and wondering how long it's going to take to get all 36 APs converted!
Here's some samples of my successes:
The syJan 1 08:09:19 udhcpc[3881]: Received SIGTERM
Sending SIGKILL to all processes.
Please stand by while rebooting the system.
[ 567.034000] Restarting system.
APBoot 1.5.5.7 (build 56398)
Built: 2016-09-08 at 14:15:55
Model: AP-205H
CPU: BCM53011/15, revision A0
I2C: ready
SKU: 2
OTP: 0xeca01028
Clock:
CPU: 800 MHz
DDR: 533 MHz
AXI: 400 MHz
APB: 200 MHz
PER: 400 MHz
DRAM: 128 MB
POST1: memory passed
SF: Detected MX25L25635E with page size 4 kB, total 32 MB
Flash: 32 MB
PCIe0: RC, link up
dev fn venID devID class rev MBAR0 MBAR1 MBAR2 MBAR3
00 00 14e4 4360 00002 03 08000004 00000000 00000000 00000000
PCIe1: RC, link up
dev fn venID devID class rev MBAR0 MBAR1 MBAR2 MBAR3
00 00 14e4 4360 00002 03 40000004 00000000 00000000 00000000
Power: 802.3af POE
Net: eth0
Radio: bcm43460#0, bcm43460#1
FIPS: passed
Hit <Enter> to stop autoboot: 0
Booting OS partition 0
Checking image @ 0x1e100000
Copying image from 0x1e100000
Image is signed; verifying checksum... passed
SHA2 Signature available
Signer Cert OK
Policy Cert OK
RSA signature verified using SHA2.
Uncompressing... done
[ 0.000000]
[ 0.000000] Aruba Networks
[ 0.000000] ArubaOS Version 8.3.0.4 (build 67604 / label #67604)
[ 0.000000] Built by p4build@pr-hpn-build07 on 2018-11-02 at 15:46:37 UTC (gcc version 4.7.2)
There's some asynchronous stuff going on there right at the reboot, as I see minor variations in logging, although it doesn't look like anything is different. Here's a few more samples of successful conversions:
Converting to Controller based AP. Will automatically reboot when done.
Th[ 303.668000] syslogd used greatest stack depth: 5084 bytes left
e systeJan 1 08:05:01 udhcpc[2263]: Received SIGTERM
Sending SIGKILL to all processes.
Please stand by while rebooting the system.
[ 308.694000] Restarting system.
20:4c:03:10:79:6c# enable
20:4c:03:10:79:6c# convert-aos-ap CAP 10.10.31.208
Converting to Controller based AP. Will automatically reboot when done.
20:4c:03:10:79:6c# [ 515.118000] wget used greatest stack depth: 5300 bytes left
Sending SIGKILL to all processes.
Please stand by while rebooting the system.
[ 628.350000] Restarting system.
20:4c:03:10:75:14# enable
20:4c:03:10:75:14# convert-aos-ap CAP 10.10.31.207
Converting to Controller based AP. Will automatically reboot when done.
Sending SIGKILL to all processes.
Please stand by while rebooting the system.
[ 292.626000] Restarting system.
20:4c:03:10:75:44# enable
20:4c:03:10:75:44# convert-aos-ap CAP 10.10.31.208
Converting to Controller based AP. Will automatically reboot when done.
Th[ 303.668000] syslogd used greatest stack depth: 5084 bytes left
e systeJan 1 08:05:01 udhcpc[2263]: Received SIGTERM
Sending SIGKILL to all processes.
Please stand by while rebooting the system.
[ 308.694000] Restarting system.
------------------------------
Cathy Fasano
------------------------------