01-04-2017 08:42 AM - edited 07-05-2021 06:18 AM
Hi all, I'm looking to do my first upgrade on my 5520 HA pair. The pair in question is for a hospital campus so, as you can imagine, uptime is crucial. The upgrade will take place at night, but I still need to minimize the downtime. My question is if I can perform a restart (soft reload) or do I need to actually do a reset (hard reload)? The documentation says to schedule a reset, but in my experience it takes quite a bit of time just to get the server bounced to even start the WLC software loading.
TIA!
Ryan
Solved! Go to Solution.
01-04-2017 10:30 AM
I have done this on many sites and with the 5520 it takes about 10min before everything comes back up. You will do a reset, but I would issue the reset system both in <time> and specify 5 min or 10 min. Also make sure you pre download the image to all AP's so reduce the downtime. One thing I would do is if you run a Microsoft DHCP server is to monitor the AP dhcp scope and verify that there are no bad addresses when you initiate the upgrade.
-Scott
*** Please rate helpful posts ***
01-16-2017 06:39 AM
Just a quick follow up, not that you need the verification Scott, but you were 100% correct. The downtime was less than 10 minutes and all came back as expected - thanks for the validation!
01-04-2017 09:37 AM
So the documentation says to schedule a reset, it means a reset not a restart as you cannot schedule a restart.
01-04-2017 09:42 AM
Thank you for your input, but the reason it says to schedule is to avoid having a version mismatch between controllers. My question is if anyone has performed a restart in the field and run into any issues.
01-04-2017 10:30 AM
I have done this on many sites and with the 5520 it takes about 10min before everything comes back up. You will do a reset, but I would issue the reset system both in <time> and specify 5 min or 10 min. Also make sure you pre download the image to all AP's so reduce the downtime. One thing I would do is if you run a Microsoft DHCP server is to monitor the AP dhcp scope and verify that there are no bad addresses when you initiate the upgrade.
-Scott
*** Please rate helpful posts ***
01-04-2017 11:48 AM
Thanks Scott, I've already got it all set with the APs pre-downloaded and images swapped. I've also gone through and issued a reset system both at <time> for my 1:00 am upgrade schedule. Again, since I've not done an upgrade in production with this configuration, I don't know what to expect from a downtime. 10 minutes is acceptable, 25 - 30 minutes (which is what I was worried it might take) is a completely different situation.
As for DHCP, we use the BlueCat IPM which has always been reliable.
Thanks again!
01-04-2017 11:59 AM
I have upgrade site with 400 some AP's and everything came back within 10 minutes. The dhcp bad address is only with Microsoft dhcp so you are good. I would have console connection to each so you can at least monitor and have a good understanding of what happens for future reference.
-Scott
*** Please rate helpful posts ***
01-16-2017 06:39 AM
Just a quick follow up, not that you need the verification Scott, but you were 100% correct. The downtime was less than 10 minutes and all came back as expected - thanks for the validation!
01-16-2017 09:06 AM
Thanks for the follow up. This helps others also when they have concerns about the time it takes for an upgrade.
-Scott
*** Please rate helpful posts ***
11-22-2017 09:46 AM
Question on the windows dhcp, we have windows dhcp server2008 and have the bad address issue, in reading I found this command(config ap dhcp release-override enable {cisco-ap | all})
as a work around and was wondering how well it works.
Dean
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide