i’m ready to pull what little hair i have left … plz help me to fix this seemingly huge consistent packet loss i am having on my rb1100hx2 … pls advise what info you need from me to help fix this " issue" i seem to be having a lot of consistent packet loss on my router lately … it gets so bad my voip and even skype drops calls, websites load slow, timeout, our websites load slow etc…
i am wondering is there is a setting we can tweak etc … some have mentioned adjusting ttl … dont know how … pls help i initially "contacted our isp thinking it was " their " network … i ran a ping monitor program and it showed packet loss on the public / upstream side of the modem … ( aka plugin computer directly to modem port / plugin ) which they seem to have fixed … but the same program now shows massive packet loss from my laptop thru the router … when i run ping test from " terminal " in mikrotik control panel it shows packet loss in a random pattern… any suggestions …
to restate … running ping program on a puter connected directly to modem we get no packet loss now .. connected TO the router we get consistent packet loss in a somewhat random sequence… aka … 10 good pings… then 2 or 3 timeouts then 12 good pings then 5 timeout etc…doesnt seem to be a pattern .. random …
’
i have a couple servers behind the router on internal ips, i have 3 public ips provided by my isp … i am using NAT… the problem, we barely even come close to utilizing our pipeline…
the packet loss seems to come in spurts… some have mentioned it might be a ttl issue ..not sure… all i know is that packet loss gets so bad that websites timeout, website pages
i am kinda new to this … i am a computer tech but this mikrotik os is not my area of expertise and know just enuf to get my self in trouble so need some advice and/or diagnosis from experts in router os configuration..
sorry for the dumb question… but wouldnt that disable dhcp … becuz i have a few devices like my printer using dhcp and ipad ??? is that what u mean by " if not needed …
unfortunately now im even worse off, my completely dead in the water, my websites i make $$$ with are not working at all and i have no idea how to fix this issue i am having to setup the 1:1 nat and until it do my websites will be offline losing all the google rank i built up … u can contact me on skype … PLZ help me!!!
Check you lan for loops as just two routers/switches without STP could bring your network down with strange effects.
Does it happen on all eth ports ?
Could you configure some ports as switch, exclude them from rules and see if it drops packets if you connect computer “directly via this switch” to modem ?
Have you resource intensive packet marking instead of connection marking ?
Maybe some filter/NAT rules “loops” packets ?
i am sorry but i have no idea what your talking about … ( except for checking cables ) which i already did … i had someone else attempt to fix it but they gave up for whatever reason so u seems to know about this …could you help me out here … work with me to get it fixed … asap .thks
Check you lan for loops as just two routers/switches without STP could bring your network down with strange effects.
If you connect two devices with two cables then you have path loop between them. Packets are send via one port and immediately return via second one. Such packets saturates devices and they cannot do anything else. To prevent such loop you need STP or RSTP protocol which “switches off” extra link.
Does it happen on all eth ports ?
Have you changed ports to check if particular one is source of long ping ?
Could you configure some ports as switch, exclude them from rules and see if it drops packets if you connect computer “directly via this switch” to modem ?
Have you resource intensive packet marking instead of connection marking ?
Maybe some filter/NAT rules “loops” packets ?
Maybe some rules sends packets back and forth and it is source of long ping.
Just follow KISS rule. Start with router set as dumb switch and add incrementally configuration changes or start removing one by one rules from current set to see which one is the culprit.
I was trying to help with this issue last week. Basically here is his problem:
He has a cable modem, in bridge modem. He is assigned a /29 subnet. x.x.x.72 is the network, .73 is gateway.
He WAS using 1-to-1 NAT and had servers on .75 and .76. Sometime in the last two weeks, his ISP made changes and is now doing static ARP for each IP. They want a single IP per MAC. The packet loss that he was seeing appears to have been caused by routing issues.
So, is there any way to get the Mikrotik to use different MAC’s for each public IP?
There is easy quick hack for this but you need one more device - simple switch
Connect modem to switch.
Assign each IP to different port inRB1100 = you will have 4 WAN ports with unique MACs..
Rember that all these RB’s ports must be separated = they have no master .
Connect WANs to switch.
Voila.
i’m sorry im a little confused here … i can easily assign an ip to each port ( in this case .75 & .76, .77 ) to each port … does it matter which one … becuz i know 11, 12 & 13<---- ( which is where " feed " from the cable modem is plugged into currently ) are supposed to have some special significance dont they over the other ports, also i’m not sure what u mean by " have no master " sorry for the seemingly “stupid” question but … i’m confused as to the logistics …
modem cable from wall to modem ------>
then u say connect modem to switch … where i’m confused .. you also say connect WAN to switch ??? huh ???
not sure what u mean by that and how will that affect my current configuration … any changes … ??
Yes.
Assign needed unique IPs to any ports you want in your RB. Connect modem to switch and all these WAN ports to same switch. Why ?
You can see here: http://i.mt.lv/routerboard/files/Block-RB1100AHx2.pdf what are “special” Eth 11, 12, 13 and how ports are grouped.
Ports 1 -5 and 6 -10 are connected to switch chips so the could be configured as 2 switch groups. Beeing connected to switch chip does not mean beeing part of switch. Let assume that all ports 6 -10 are part of switch. It means that one of these ports is master for rest. Let it be port 6. All others are slaves. If you connect cable to any port of this switch it will have MAC address of master port, it means of port 6. Letter “S” mean “Slave”. There is “S” for Eth6 as it is slave of BRIDGE but Eth6 is master for Eth 7 - 10. Ports 1 - 5 are independent “singles”.
See the picture.
masterswitch.png
Let assume that it is not a problem now and use any port you want but check if they are not part of switch
ISP’s modem carries a few IPs (subnet) via one eth port. If it was possible in Mikrotik then you would assign set of unique IP and MAC pairs to one ETH port and connect this virtual ports directly to modem. But it is not doable.
Therefore we need to assign IPs to unique ports to have unique MACs. Then all that ports we need to connect to modem.
The only way is to connect them to some kind of concentrator/funell which “squeezes” it to one port and “unsqueezes” subnet cominig from modem to these WAN ports. Switch is such funell.
i initially "contacted our isp thinking it was " their " network … i ran a ping monitor program and it showed packet loss on the public / upstream side of the modem … ( aka plugin computer directly to modem port / plugin ) which they seem to have fixed … but the same program now shows massive packet loss from my laptop thru the router … when i run ping test from " terminal " in mikrotik control panel it shows packet loss in a random pattern… any suggestions …
