04-24-2015 08:12 AM
We have been seeing issues with users getting dropped from our wireless network. We are a mixed Windows / Mac envirnoment with both clients seeing this problem. We recently switched out of a mixed 802.11n and ac environment. We are now running only 225 access points in AC mode.
By going to ac only we have seen performance improvements, but the issue with dropped connections remain a major problem. I recently spoke to a user who had wireless disconnects 4 times in an hour long training session.
All help is appreciated.
04-24-2015 08:21 AM
points in the cluster
*Jim McMurchie* | *Smartsheet*
Winner : Best 100 Companies to Work For 2014
04-24-2015 08:40 AM
I would recommend you do the following :
- Disable 80 Mhz channels under the RF/ARM settings
- Enable broadcast filtering "All" under the SSIDs/Advanced settings
See if this helps you.
A couple of other questions:
- What type of authentication are you using ?
- InstantOS version ?
Lead Mobility Engineer @ Integration Partners
AMFX | ACMX | ACDX | ACCX | CWAP | CWDP | CWNA
05-06-2015 10:42 AM
I'm not trying to hijack this thread, but I'm having a similar issue with users being dropped at random times. It also seems that once it happens to one, it happens to at least a few of them on the same cluster, but connected to different AP's. We also have the IAP-225, but we have 61 in our cluster. When I'm looking at the virtual controller, I'm not seeing anything wrong when the users are reporting outages. I will sometimes get alerts that a DHCP request has timed out and occasionally a RADIUS timeout, but that's fairly typical throughout the day and not limited to when users are experiencing issues (and it's typically not reporting an issue for those clients who have contacted me). Over the 61 AP's, we're typically running in the range of 300-350 clients, the vast majority are Windows, but there are more than a handful of phones/tablets. We're broadcasting out 3 networks, our secured network and 2 guests. The clients we get reports from are on the secured network, but then most of them are connected to that network (I'd say 90% of them).
I did try the suggestion to turn off 80Mhz channels and Enabling Broadcast Filtering, but it doesn't seem to be helping. I'm willing to post my config if you think it's a setting.
Thanks for any assistance!
Wellsboro Area School District
05-09-2015 08:04 PM
This is probably not your case, BUT...I had an admin at one point who had installed an Instant cluster on VLAN 99 as the native VLAN, then allowed wireless clients to get DHCP on the same subnet - VLAN 99.
I would've thought this would not work AT ALL, but it worked long enough for the customer to be convinced he'd configured it correctly. Two weeks later, all Hades broke loose and half of his clients at any given time were falling off the network.
This may not pertain to your config (and certainly not if TAC has already looked at it) but may remind someone else to check the obvious!
05-11-2015 10:09 AM
Wanted to update the thread with the changes made (by Aruba, Samya is awesome) that seem to have fixed the problem.
1. Reset the 2.4 Ghz channels to standard 1, 6 and 11
2. Lower power output to 6-12
3. remove 80GHz support
4. Disable 2.4 GHz radios completely.
It has been over a week since we put in these configuration changes and we have only one report of anyone being dropped by the network. Once the dropped connections were sorted, we were able to map two physical areas of poor performance. In each case we had a poorly performing 225 IAP. We had sufficient converage to shut down one of the two units, the other we replaced with a spare 115 we had in stock.
05-11-2015 11:14 AM
I've attached my config. We now believe the users are being dropped because of a wired issue that we are having checked out that's also affecting the AP's.
Like the OP, we dropped our minimum transmit power a few weeks ago and, along with the broadcast filtering and disabling 80Mhz channels, it did seem to help. We can't disable 2.4Ghz as we still have some equipment that only have 2.4Ghz radios. We thought that misconfigured rSTP on our wired switches was the culprit Friday, but alas, it has since continued to go down randomly. So maybe it's a mixture of issues.
Please let me know if you see something amiss.
Wellsboro Area School District