I have some trouble with my RB-493G's ethernet ports. The problem occurs random times (sometimes 3 times in a day, sometimes 1 time in a week). There is no rx packets on one or more ethernet ports. It is totally random which ports are in trouble. First sight, all is well because the interface has R flag (status is 100 or 1000 Mbps / full duplex), but only the tx packets are incrementing. The connected devices to these ports are various Router Boards (RB-750G, RB-411AH, RB-433AH), which are shows similar symptoms in this case (linking eth ports, but no rx packets, only tx). The issue can be solved only with the /system/reboot on the RB-493G. All the 9 ethernet interfaces and wlan1 (R52Hn) are members of a RSTP bridge. I've tried it with 5.6, 5.7 and finally the 5.8 ros versions.

Has anybody similar issue with this type of board, or I've got a defected one?

I have some trouble with my RB-493G's ethernet ports. The problem occurs random times (sometimes 3 times in a day, sometimes 1 time in a week). There is no rx packets on one or more ethernet ports. It is totally random which ports are in trouble. First sight, all is well because the interface has R flag (status is 100 or 1000 Mbps / full duplex), but only the tx packets are incrementing. The connected devices to these ports are various Router Boards (RB-750G, RB-411AH, RB-433AH), which are shows similar symptoms in this case (linking eth ports, but no rx packets, only tx). The issue can be solved only with the /system/reboot on the RB-493G. All the 9 ethernet interfaces and wlan1 (R52Hn) are members of a RSTP bridge. I've tried it with 5.6, 5.7 and finally the 5.8 ros versions.

Has anybody similar issue with this type of board, or I've got a defected one?

Sounds like a hardware issue. Swap out the power supply to eliminate that. Still bad after that, consider replacing the unit or warranty repair if applicable.

What packages are you running on the 493

What packages are you running on the 493

I don't know of any particular packages on the 493G that would cause this problem.

What packages are you running on the 493

I don't know of any particular packages on the 493G that would cause this problem.

Not sure if this is relevant to the post but from my experience http://forum.mikrotik.com/viewtopic.php?f=2&t=56080, now i disable unused packages and any disabled setting but it maybe as you have suggested a faulty board.

I've changed the power supply yet, but experiencing same symptoms unfortunately.
I'm using these packages on RB493G:
advanced-tools
routerboard
routing
security
system
wireless
All the others have disabled. I will change this board to another RB493G today afternoon.

It's unbelievable! I've changed yesterday the original RB493G to another one and I've reconfigured this new board from default (empty) settings. The ethernet lockup problem still exist. What the hell is going on? I've no idea how can I solve this...

What does the log say on these boards?

I haven't found absolutely nothing in the log yet. But only the critical, error, info and warning topics had been switched on at the /system/logging. What topic(s) is(are) relevant in this case? Interface? I've just enabled it.

Was the board working OK and only recently this problem occured or has this issue been happening since day1?

I have experienced this problem since when I changed the original set (RB411AH+Allied Telesyn AT8024 L2 switch) into these RB493G's.

Ok - since you changed over to 493 this issue happened, i would suggest you post a diagram of your network and where the 493 is located in it. Is your network bridged if yes i would suggest a switch to routed and if you using the 493 to simply bridge to the other MT devices why not use the AT8024 for that as performance of a MT device is reduced when software bridging is used as opposed to hardware bridging.

Ok - since you changed over to 493 this issue happened, i would suggest you post a diagram of your network and where the 493 is located in it. Is your network bridged if yes i would suggest a switch to routed and if you using the 493 to simply bridge to the other MT devices why not use the AT8024 for that as performance of a MT device is reduced when software bridging is used as opposed to hardware bridging.

Software bridge, while it may lower performance, will not cause an interface to lock up. Perhaps whatever device is plugged into the interface is sending ESD or some electrical signals out of spec.

I'm not alone with this issue. I think, I've exactly the same problem like described here:
http://forum.mikrotik.com/viewtopic.php ... 01#p289601
I've tried two RB493G's and two RB435G's and the ethernet port freezes always appearing. Currently the RB435G's are in use. Both of them have ROS 5.8, one of them has firmware version 2.29, the other has fw 2.36 (the latter more often occurs in the problem).

Ok, did you make a support request to MT? This issue is all over the forum but MT still denies great part of it. They say nobody makes report of Ethernet ports failing or flapping. So please send these reports in!

Make supout.rif's of all connected units. State the software version, and firmware.
Also make a network diagram and tell us what cables are used; cat5e or higher, shielded or unshielded and in case shielded, indeed with shielded connectors?
Also, are are units properly connected to earth (grounded)?
Do you see the interfaces switch back to 100 or even 10Mbps at times?
This is all important information.

I would like everybody to make support request to MT as much as possible when it goes about Ethernet issues.

Hello Wirelessrudy,

I've read your posts for 2 or 3 weeks ago, but I think that my problem might be different than yours. If I remember correctly, you've described that your eth interfaces are flapping. Does it mean the data flow is interrupted and then continues on its own, doesn't it? But in my case, the data flowing completely stop (the eth interfaces freeze - no rx packet received anymore), and only the reboot solves this problem.

I haven't made a support request yet, because I've believed that, it might be a hardware problem especially with RB493G's. But since when I've changed RB493G to another one, then changed them into RB435G's (and it couldn't solve this problem), I've searched the forum for similar sympthoms. I've found SXMU's post (http://forum.mikrotik.com/viewtopic.php ... 50#p285372), that is exactly the same problem than mine. Do you think your problem is that the same?

One more thing what I noticed: when I was using the RB493G the eth port freezing concerned either the 1,6,7,8,9 ports, or 2,3,4,5 ports. RB493G has 2 switch chips, the switch1 handles, the 1,6,7,8,9 ports, and the switch2 handles the rest. It never occurred, that all make mistakes, always one or the other interface-group were in trouble. Since when I've used RB435G, which has only one switch chip, ofcourse all the 3 ports are in troble, when the problem occur.

I have similar problem
http://forum.mikrotik.com/viewtopic.php ... ex#p302291

went to this forum to ask about similar forum and whoa!
Can confirm. RB493G. Current version 5.13, still problems.
LED blinking, interface is active in panel, but no data and no DHCP lease. Solved only by removing power cord and inserting back (software reboot works from time to time)

I can see similar issue after I upgraded from 3.30 to 5.14. Is this solved somehow already? Some driver sensibility was changed or what is going here? My customer start complaining. I through, that it is his PC issue, so I connected my laptop. And can't get IP address through DHCP. So, I started WireShark, and I was able to see CDP packet coming from RouterBoard. But was not able to see DHCP packet on RouterBoard coming from laptop. But link up on 100M/full. I changed manualy speed on my laptop to 10M/full and I got IP from DHCP. After that I unplug cable from my (manualy 10M/full) laptop and connect it back to customer PC (auto/auto) and his PC got IP through DHCP, too. I am not sure where is problem here.

Do somebody know solution? Now running on ROS 5.15 and FW2.39 on RB493AH. But still same problem occurring from time to time. There was NOT this problem on ROS 3.30 before. I must downgrade.

May be basic idea, but did you try to change UTP cable for sure? I have seen twice that one component works fine, UTP was checked by simple tester like this. And another component didn't work well till I changed UTP for new or use second pair for data. For sure use check link using another UTP min length 1m.
I'm actually running on RB493G ROS v 5.11 and everything seems to be OK.

Already downgraded back to ROS 3.30 and FW 2.20, I will inform you later, if downgrade help or not.

We will get a solution ? coz it's annoying.

Try the bios firmware 2.41.
Today I flashed 493g, for the past 6 o'clock no drop interfaces. Yesterday, during the same period there were two falls.
Hopefully this will solve the problem.

ps: Sorry for my bad English.

fbi or anyone else SEE if its something similar like this posts

http://forum.mikrotik.com/viewtopic.php?f=3&t=25068
http://forum.mikrotik.com/viewtopic.php?f=3&t=61574

I m waiting this to be fix long time

Anybody received an answer for this? i'm still waiting for an answer

Ticket#2012073166000444

Bump!

Any news about this issue?

Don't know if I can help, but I would like to know what power supplies you are using on these devices.

If you are using a 24v power supply, have you tried a power supply with a lower voltage rating, like 18v or 12v, just as a test?

I try with 24v Poe and Jack and with 18v, latest routerboard firmware and latest ROS version, no luck with that

Hello,

I've been on the same trouble for almost a year using a 493AH, that I tried to replace with a 493G with no success (instead it went worse, so I reinstalled the 493AH). On the other end I had 3 RB433AH. No matter in which port 2-7 I had them connected. I also installed new cabling from the same provider I had already installed.

Well, I had two issues:
- Permanent: no 100mpbs full-duplex link using about 3 meter of supposed cat-5e/cat6 cable. Link sync changing constantly between 100FD right after reboot, 100 Half a minutes later and even 10mbps!! (FD/HD depending on unknown conditions).
- intermitent: loose bidirectional connectivity on the ethernet.
- once: total loose of connectivity, even with electrical reboot of the boards.

Last week I replaced everything for a couple of 450G and nothing changed, still had the first syntom: 100 half duplex.

Then I changed the cabling and connectors to two RB433. I used an apparently similar cable but from different provider and metallic UBNT connectors with the ground wire attached on both ends.

Everything worked like a charm since then. I still have to change one cable.


The cable I used was 4 twisted pair, 24AWG solid copper with ground wire.

Try using a different cable

Then I changed the cabling and connectors to two RB433... Everything worked like a charm since then. I still have to change one cable... Try using a different cable

And invest in a Cat5 cable tester! Even one of the inexpensive ones (~$10) will save you some big headaches.

This problem is fixed in RouterOS v6.x