[ietf81-tech] Access point reboots

[Bill Fenner]

I'm running a script on the management box that polls all the APs every 10
minutes, and an AP has no clients associated with it, and if it hasn't done
so yet, it reboots the AP.

This is to get around the recently-discovered mode-resetting problem: the
config doesn't get saved, or fully used.  I've previously explicitly saved
the configs on each AP, and chelliot has updated the autoprovisioning code
to work around this failure, but the APs that are out there and running
could lose their IP address at any time due to the weird DHCP behavior.

The code is on the management box in ~fenner/src/pupu/reload-idle-aps/ .
 The list of APs being looked at is in aps.txt; the list of APs that have
been reloaded is in aps-reloaded.txt; the script's output is in loop.log.
 You can find out what APs have not yet been rebooted by the script with

[fenner at management ~/src/pupu/reload-idle-aps]$ comm -23 aps.txt
apnames-reloaded.txt


The current list:

Still to be reloaded at Sun Jul 24 15:53:28 GMT 2011:
ap101 ap102 ap103 ap104 ap105 ap110 ap111 ap112 ap119 ap120 ap121
ap123 ap124 ap125 ap127 ap130 ap131 ap134 ap138 ap139 ap141 ap142
ap144 ap145 ap147 ap151 ap153 ap157 ap159

(note that this list may have un-deployed APs in it; when it gets down to
only un-deployed APs I'll stop running the script.)

  Bill
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.daedelus.com/pipermail/ietf81-tech/attachments/20110724/93174746/attachment.html