yes, we have a lot of them is without the issue since a day one, but when the one having this issue, the only way we can do is, just replace it instead of waiting time to investigate.
just can say that, in 100 board, there may 5~7 unit have this issue, which all the configuration is the same and all is the RB433. suspect, may possible due to the QC issue.
just worry that, seem mikrotik customer getting many and many, the QC may start drop and it creating more and more hardware problem… hope it will not, but… we facing it now…
I have the exact issues as you guys do. It is random across the network. I tried a veriety of different versions. Right now I am on 4.11 and the latest boot firmware. I am creating supouts for mikrotik. They dont know what to do. I disable and reenable the ethernet port and it starts working. Otherwise I dont get any recieve packets. As of right now. My network is offline. I cant macping. I just disabled and enabled and it started working. I noticed only the routerboard 433 so far keep locking up. I dont recall any other version having the problem as of yet. This is really anoying. I cant stay up every other night disabling and reenabling. This needs to be fixed now. I cant do business like this.
I don’t understand why MT is not even letting us know they are working on it!
Some of you guys tell us they even notified support but still no answer from them??
I think we all need to start writing to MT because this problem is becoming very, very anoying!
When will my rb’s again freeze???
My business model is based on doing “better” and “more reliable” then the competition. But this way I can’t fulfill my promise to my customers…
MT cannot publicly acknowledge that there is even the possibility of a problem.
If it was a Manufacturing or QC problem, then MT would already have said so and recalled/replaced the boards.
It is much more likely that it is an Atheros Chip fault and/or coding fault originating with Atheros.
MT is bound by Non-Disclosure Agreements with Atheros, and that stops them saying anything.
Remember that the 400 + 700 series and all their radio cards are Atheros based, so if MT even hinted at pissing Atheros off by saying there might be a hardware problem, then they would be in deep doo-doo.
Ubiquiti found a workaround, and so did Mikrotik. MT’s workaround seems to fail noticeably. Mikrotik are probably (hopefully) working on a new workaround for what is probably a chip fab error, but they cannot ask for help from anyone but Atheros.
It is likely that Ubiquiti signed another NDA that stops Atheros giving their solution to MT, unless they pay lots of money for it. A Microsoft-inspired ‘reboot it as often as needed’ solution isn’t such a brilliant workaround though.
It is a real shame that all of the Skill that is out there cannot be brought to bear on the problem, because Atheros has a policy of telling Nobody Anything, despite the popularity of their products.
Basically it will get fixed silently in a ROS update in the near future (i hope).
(watch for Normis asking hard about what ROS version you got)
If you experience problem where RouterOS reports link as running, but traffic cannot be received, try to unplug Ethernet cable. If Link/Act and 10/100 LEDs are still on then there are problems with Ethernet PHY.
adrianatkins:
I doubt if this has anything to do with Atheros. I have the issue with my rb1000 that has no radio cards in it.
I have two rb411’s without radio cards that still have the issue.
kirshteins: If the ethercable gets unplugged the ROS shows “no link”, plug it back in link comes up and even speed and duplex/simplex negotiation takes place.
So all parameters show the link is up but still no recieve traffic.
In some occasions I was still able to do a mac ping from a healthy unit to the problem one but noting more. Not even a mac telnet session.
Some units though didn’t even allow mac ping…
Once I logged into the rb1000 through another port and did a reboot which then didn’t bring the port back alive.
After a power cycle the port came back and worked fine…
That routerboard 1000 is now running for 4 weeks already without hickups but even writing this makes me fear…
I got many troble from LAN connected customers if theirs PC not grounded. It have some 120V AC on PC case ( I have 230AC electricity ) and same voltage on other end of cable connected to my SOHO swich. I restart swichs 1-2 times per week until I make grounding for each PC That’s is big puzzle for most LAN ISP
Well I agree. My suspicions are also in that direction. I have been told that if you use shielded cables both end connected devices need grounding anyway otherwise the signal migt become worse instead of better then usage of normal unshielded (and thus 100Mb connection) cables.
So I started doing this and actually in the first time in my live as a WISP started also to ground all routerboards and antennas since I lost several radio’s this spring due what I only can think of, static electricity build up.
But in doing so I ran into more problems. According a qualified elec. I also needed to ground battery pack and solar panels in the instances where they were used. And ALL devices, racks, boards, antenna’s MUST have their own ´star connected´ ground cable.
He found 3,5V up to 12V potencials between devices all running on 24V. I actually had some shortcust untill we found that batteries and solar panels with their controllers are to be grounded to + where all other equipments and antenna’s etc to be grounded on -
He than also found that the rb1000 rackcase are not CE approved. They lack a groundi connection and according proper electric enginering rules the supplied 220V to 24 (or12v?) adapter is not supplied with any ground neither. They should if they would have been European CE approved.
The problem with the rb1000 case is also that they are powder coated, like the rack chassis they are in, and that powder coat serves as insolator.
So what we have done is scratch the rb1000 case and the rack fittings to connect bare metal to the earth wire.
We have done for all devices in the rack and some rb’s loose in the box and the batteries are removed from ground since that obviously did conflict with the ground of the routerboards.
To be in short, after this groundings we have the rb1000 running without hickups…so far
I have to see in the long term, 4 weeks is not so long.
Also, to revert to my earlier posts; the rb1000 having that etherport problem has been running for 3 years continous on a wooden pallet connected to a loadbalancer on one end and a switch with several rb’s on the other ports.
No grounding, just unshielded cables. Never had the problems!
(running ros 2.xx at that time…)
I have run into the same isue with the RB433’s. I use Engenius 2610’s and ran a LAN cable between its port and the ethernet port 2 on the RB. Ethernet one (which also has the POE) has a customers computer plugged into the AC Brick. Ports 1 and 2 have no receive but good transmit. RB’s etherent port 3 goes to a Cisco 2960 and is the main path going back to a RB1000 and I have no issues. For ports 1 & 2, I show a matching Tx rate on both. Disabling/enabling didn’t help. Part of the day I can ping the Engenius that is attached to port 2 but then I lose it. Even when I can ping it I see only Tx traffic on port 2. I just ried rebooting the RB and it didn’t change anything. Even rebooting the Engenius doesn’t help. If by chance maybe I have a setting wrong, maybe someone could let me know.
Let me think back to the past. My ethernet ports were locking up randomly. Wireless and routeros was working but ethernet was crashing. I kept trying different versions of software. 3.28 it stopped. Just recently I upgraded all my rb411s to 5.0rc10 and I do not have any ethernet issues. However I think your issue is probably one of these things. Please try some of them.
Replace cable with new rj45 ends
Upgraded to the very latest 5.0 and make sure the boot loader is the latest. I think 5.0 automaticly does this. You can check.
Very important. Make sure no power source/power lines are within a foot of the ethernet cable. There is a EMF field. Based on the length of the cable and the amount of electricy going through the cable. It will cause you ethernet issues.
After all those options if you still have the issue. It is possible the internet port is bad. That can happen if in the past it got a surge from the power lines. But I think if you do the list I gave you. You will get it working.
