This is why I dislike UniFi switches

preview_player
Показать описание
My primary switch at the house which is a UniFi switch completely locked up and was only partially passing packets but what’s worse is that it wouldn’t change any of the switch port states. I’ve seen this before and it’s happened in media production environments where it seems that be busy box software and the CPU for years but the switch ASICS keep switching. In my personal opinion this is unacceptable for a switch to do this. If a switch locks up there should be some sort of watchguard that forcefully reboot the switch. Keep an eye out for another video where I replace my UniFi switches with some other vendor. And for those actually reading this description the hard power cycle of the switch works to bring it back online.

Рекомендации по теме
Комментарии
Автор

Had this happen once at a site. In my instance, two keystone Jacks that were side by side had copper from the cat6 extruding the jack and touching each other. Not saying that's the problem here, but if you have a tight fit on the keystone's, might be worth a look

derekg
Автор

For production environments use the Ubiquiti EdgeSwitch line and not Unifi. Their Software Defined Network products are not meant to be in heavy production environments. The Edge products have much more functionality anyway and are much more reliable.

algheroman
Автор

For the cost of Cisco, you could probably replace that faulty switch 5-10 times at least and still save money lol

estusflask
Автор

I've seen this before on cisco enterprise switches while they are locked up.

Honestly it doesn't matter what your link lights are doing when a switch is locked up and not forwarding traffic, it's just time to pull the power and hope your flash & config are ok.

joshmonus
Автор

As an network enthusiast I deployed 3 Unifi switches and 1 Edge Switch at my office and at home. It has proven to me that these switches won't stand the test of time, with all of them failed after a few years. The very issue they have is the consistent powersupply failures, which could be caused by the weak heat dissipation design. Moreover, the electronic side of the switch is often so weak that it is easily fried by a faulty powersupply. This makes the switch very hard to repair. I ended up buying used Cisco Catalyst 3750X-UL switches for replacement. They worked like a charm, very steady performance and trouble free. The very best side of the Cisco is that the switch use hot swappable power supply. If the power fails you put on another power module and you are back in business. Of course, one has to have some understandings of CISCO IOS before using a Catalyst.

If any one can overcome the intial learning curve for the CISCO command lines, I would strongly recommend replace all the UBNT switches to CISCO ones. Those switches are designed by professionals for professionals. Don't waste your time and money on UBNT switches.

catttcattt
Автор

Had it occur 2 times within a period of almost 2 years. As this is within a home lab, it is not that bad (just a couple of kids freaking out when I reboot the switch mid day in a weekend). Indeed unacceptable for production environments. What worries me though is that not only the led's keep blinking, but in my case, as long as the patch was in the port, traffic kept flowing. Removing and reseating the patch stopped the traffic and I needed a reboot to make it work again. Apart from this, I'm quite happy with unifi gear (but as said: this is for home lab)

realburn
Автор

I've had this happen a number of times on HP ProCurve switches (in fact most switch failures I've dealt with were ProCurve and I don't even manage very many of them!) but only once on a Unifi switch, and I manage multiple hundreds of them, some in very large deployments with a LOT of cameras, POE, and traffic going on, no problem. One switch that had this happen had been damaged by a network cable getting shorted while actively delivering POE....and that's not the switch's fault. I run a pair of 16 port Gen1 POE in my home environment and have never seen this happen, and I create a lot of traffic on a regular basis, along with several cameras, waps, NVR, etc, and all SFP slots in use.

pawo
Автор

Interesting...I've never had this problem on my UniFI gear. Did you ever find a cause or solution to the problem?

Naure
Автор

Your switch is broken, I've never seen this on dozens of unifi devices.

ukrailway
Автор

+1 couldn't agree more -- I have a couple of UAP-AC-PROs connected up to a gigabit poe Cisco Switch SG250-08HP -- Cisco makes rock solid hardware even if the software doesn't look as pretty, it works soo nicely!

fossjonc
Автор

My issues with UBNT,
1. They will discontinue a product without telling everyone or publishing EOL. They just go "oh we ran out" we are not going to restock it.
2. for the same product sku the item can be different. For example, all the "AC-PRO" they used the same product code instead of putting verison2.0 when they switched from the Broadcom chip set.
2a. unlike cisco, where they put HW verison vs v1 vs v2 and etc...
3. they don't keep their old products specs online under some EOL life page. So when they "refresh a product" it's difficult to tell the difference".
4. For the unifi switches, at one point they had the option for dual PSU. Now even with the SAME part number, it only comes with 1 PSU and you HAVE to use their UBNT UPS which they never have in stock.
5. They will release a product/ new ecosystem then just randomly without warning just discontinue it. I started migrating to their light switch then BAM discontinued.
6. basically no support, their support is reddit.
7. their routers/"security gateway" has random packet loss and drops that I still haven't figured out gave up and replaced with a Palo Alto.

with that said, the only thing I would deploy with UBNT is their AP and MAYBE EDGE switches for edge ports. I would never use them for server switches or core/dist switches.

potato
Автор

That happens to Cisco, HP, Dell, Aruba, <insert_branname> "cheap" or first tier switches too. Unifi is in that low tier, so that is expected once in a while. When this happens, it is usually because the board switch is a lemon and the batch from which it came from it also plagued with the same problems. Or, the switch "cooked" too long at high temps without proper cooling.

If you don't want to see that happens, you need to fork at least 2k+ per switch no less.

Edit: And people saying Cisco SG line is good : you must not be working in networking so say such false statement. SG is the uber crap line (even below Meraki which is already quite at the bottom) of Cisco. If you want to talk about Cisco quality switch, you need to start at Catalyst and go up. Anything else is just garbage.

Traumatree
Автор

This has happened with us with Juniper. Something has probably been not properly grounded and has gotten hit. Lord forbid it be your fault because that would be “unexceptionable” 😂

Itsoperated
Автор

I have not seen that with a UniFi switch, however, I have seen that on a consumer grade NetGear switch..

silverstangs
Автор

For the price & ease of use. Let’s me focus on other tasks. Besides a simple LED issue, how stable has it been?

Multicast is not perfect & a reason I push to VLAN the crap out of things

AMPTechGrade
Автор

Maybe do a firmware update in a while?

BasicITStuff
Автор

This has never happened on my USW-48-POE switch. Sounds like you got a bad apple. I current own a UXG-Pro and a USW-48-POE and they update their port statuses accordingly

phillipkujawa
Автор

We have 9 UniFi Switches, we are not experiencing this anomaly.

keithbateman
Автор

I've been working with Ubiquiti switches for the last 2-3 years in enterprise settings and have run into so many bugs, many of them minor annoyances but some of them are horrible. I get nervous anytime I have to work with these, the software is so cheap and fragile.

largebeppo
Автор

I had a stuck port on a 48 port Ubiquiti that actually stayed stuck across reboots. The link LED was on, but no connection. I was sure the port had died. Nope. It was a known firmware issue. I found some fix online somewhere and it worked again. I think it's been patched in firmware now.

joels