Dear All … I’m opening a real challenge that consumed my health , my time , and not only that … my clients are annoyed like crazy and i’m loosing my job … (Yeah it’s real )
The story begins and ends at two times of the day
2-4 pm
11:30 pm - 3:30 am next day
The picture below will explain how my network is ..
Notes :
(1) some clients have Access points acting as wirless APs in their homes because they are using labs
(2) some clients have switches at their homes to serve more than one computer
(3) I work with hotspot and IP config 10.0.0.2-10.0.2.254 (750IPs) Gateway 10.0.0.1
(4) I have a local file sharing server for data sharing (minimizing download)
It’s actually a big lan for serving the internet to people having no ADSL service available at their locations.
At these times .. i find internet getting too slow to people specially behind nanostations and little slow at people who are on the physical medias .. but in general the network goes crazy
Pinging remote nanos(offcourse when i’m on main switch) gives very high latencies 200ms-500ms and even too much request timed outs (Should be 1ms-2ms)
pinging to near nanos gives 12-20ms (also abnormal)
pinging clients on the near lan gives requests and high latencies too
When i go at location of RB250 (2) i find the LEDS are blinking like crazy not only in this switch but also at all switches connected to the network .
I tried to disable every possible path to see if the problem resolves ..
Some times when i disable path (b) the loop stops some times when(A).. and sometimes when i disable the cable going to RB250(2) it stops .. (THEN IT COMES BACK ! ! ! )
Some times when i restart the switches it stops ( THEN IT COMES BACK ! ! !)
No rules you get me ?\
It feels like there’s a ghost in my network that insists to kill it
Any help guys ?
Please for God sake i’m really loosing my clients ..
Inspect the traffic on your network (port mirroring, wireshark) to see what is beeing transfered arround, also sounds like mac flooding attack http://en.wikipedia.org/wiki/MAC_flooding
For a network this size you should really go for routing instead of bridging.
You have given a diagram but without configuration details from each MT device.
What does “another 30 switch connected sequentially” mean?
“(1) some clients have Access points acting as wirless APs in their homes because they are using labs”
Please explain, also has these clients a CPE, who is registering into these AP’s are they also on your network?
Have you done a IP scan to check for a duplicate IP’s on the network
From your network diagram it is not possible to determine the source of the loop. Try to limit broadcast storm with broadcast storm control feature:
http://wiki.mikrotik.com/wiki/SwOS#Forwarding_Tab
I also have problems with loops. But my setup is very simple and was working, before i used RB250GS! I think, there is a problem. Also a friend told me, that his RB250GS was stop working after few days.
My Network-Setup:
HP ProCurve 1700-8G
PORTS
| | | | | | | |
1 2 3 4 5 6 7 8
V1 V1 V2 V2 V2 V2 T2,1520 T1,2,1520
| |
| \
V1520 V2 V2 V2 T2,1520 \- LINUX ROUTER
1 2 3 4 5
| | | | |
PORTS
RB250GS
Tx = Trunk-Port
Vx = Port-VLAN
Port 7 on [HP ProCurve 1700-8G] and port 5 on [RB250GS] are connected and configured as trunk-port for VLAN.2 and VLAN.1520.
After running for 1 day and a half, RB250GS is stop working. No response on pings and the switch does not forwarding any packet any more. But there is no source, where a loop can start. After replacing RB250GS and using a TL-SG3210, all worked fine. Currently RB250GS is only used as a switch without VLAN with capability to monitor the throughput and enable / disable ports.
So i think RB250GS is to cheap to get used as a replace device for a semi-professional devices as TP-Link or HP Procurve Switches.