I will setup a test router sometime next week and let you know but even if this works, I don’t really feel comfortable using betas in production routers. Shouldn’t the fix also apply to future v4 releases?
The v5.0beta5 seems to work as it should with the default settings (auto negotiation turned on at both sides). However, I still get errors and extremely slow connection problems when I turn auto negotiation off at both sides and I set duplex to half at 100Mbps. I have tested with a couple of different switches and every time I set duplex to half I get the same bad results.
But wouldn’t you expect extremely bad throughput on half duplex?
Why do you need to use half duplex at all on wired Ethernet ports?
Yes, I would expect slower speeds but not EXTREMELY BAD.
I don’t need half duplex in case I have full. However, there are still cases of people having hubs which means half duplex or even faulty auto negotiation which also usually means half duplex. I was under the impression that 1100 is suppose to work in half duplex mode also, or is it not?
Fair. What kind of results are you getting?
Something like 100-200k!!!
After some more extensive tests, even the beta is not working right. All interfaces seem to open at 100/FDX with auto negotiation on however the transfer speed is really slow again no matter how many times I recycled the Cisco switches or the 1100s. THE ONLY way I got the 1100s to work correctly even with the v5 beta is the same as with the v4 and that is by turning auto negotiation off everywhere (Ether11-13) and thats all there is to it! I don’t think this can be fixed by software upgrade as IMHO indicates a hardware problem. I wonder what Mikrotik has to say about all this though.
Please send supout.rif file to support@mikrotik.com. Generate it while you have auto-negotiation enabled on ether11-ether13. What exact devices are connected to those ports?
This sounds like standard Cisco auto-negotiation problems to me.
Just lock the ports each end at 100mbit FDX and you will have no problems. This is not limited to Cisco+Mikrotik, it also happens when you try and connect many other brands of equipment to Cisco switches.
Please list exact device names you are experiencing problems with when connected to RB1100 ether11-ether13. Also state their transmission parameters (speed, duplex mode) if auto negotiation is disabled.
All 3 (11-13) trouble making ports are connected with a couple of Cisco L3 2948G switches. The only way that worked without giving me trouble was with auto negotiation off at 100/FDX. I will be submitting the supout file soon.
I am not so sure about this here. Some early Cisco products (and a few other brands to be fair) had these auto negotiation problems as you say but I am afraid this is not the case now. In our L3 2948G switches we have connected just about anything from 3Com, Compex, Intel, other Cisco, Broadcoms, variety of Realtek chipsets and the only ports that have manual duplex and speed are those connected to the 1100s. If I lock them everything works fine. However, just for testing purposes as I also mentioned before, I locked both sides at 100/HDX and they still had problems.
It seems to me that the pressure to send new products out is causing big corners to be cut.
As ROS on Mikrotik is kind of Mature, we expect it to work.
That is a compelling reason to Buy Mikrotik - we believe it will work out of the box.
IF i have doubts that the MT New Product might Not work, and therefore do not Buy It until the bugs are ironed out, only Mikrotik Sales are affected : we can buy other stuff.
I cannot afford, nor do i want, to spend my money and Wait for a Fix for equipment that doesn’t work as expected - that affects my Money, my Service, my Customers and my Reputation.
That’s the main reason i fell out with Ubiquiti as regards AP equipment - their CPE is great VFM when you can get it.
Feeling that MikroTik is starting to make a Big Mistake right now is disturbing, and not just as regards the RB1100.
I sincerely hope that MT can be honest with their Loyal Customers, and sort out the Plan.
My hope is that MT keep on making great kit at affordable prices - not cheap, just Good, Affordable and Working 99% (nobody’s Perfect).
Paradyne ADSL2 modem, 100/full duplex. Works on Ports 1-10, but every 10-30 seconds, goes offline on 11-13. D-link DES-3028P.
Tried on 5.0b5,6,7 w/ latest boot loader.
EDIT: Both 100 and 1000 ports on Dlink. Also tried forcing 100 FD and 10 FD with same results.
We are having trouble with our RB1100 's , Ether 2 connected to a cisco 3500 switch drops every 10 seconds or so when speed is fixed to 100 and duplex to full, when both ends are set to auto it can work fine for 3 or 4 days then the port will keep dropping randomly ! This of course is proving to be a nightmare with ospf etc etc.
Is there no chance of Mikrotik continuing production of the RB1000 ? WE had no problem with them, i have tried to get old stock everywhere but no luck !
Regards
Darren
Have anyone tried to reinstall problematic RB1100 using Netinstall and reproduce this problem with default configuration?
I think there appear to be many people having the same problem that it is most likely hardware, our config was manually entered, it still had this problem with only a single ip and default route set.
Regards
Darren
I had the same problem I cant reintsall via netinstall because is operating.. only in weekends can test it.. please mikrotik lets us a solution
ok guys, rb1100 into cheap rubish switch and it has not dropped now for a week ! it just seems to be with cisco’s !
Regards
Darren
What relevance would that have. You have A LOT of people having the same problem. Default config or no, the device wont keep a link with many different types of hardware…
We tested multiple RB1100 with Cisco L3 2948G and did not experience any problems. Please describe what kind of traffic is being forwarded, configuration of both devices and other conditions to reproduce this problem.