Background:
I bought a hAP ax³ AX Home Access Point 2.5 Gigabit - Router and it´s a great router.
However after powering on after cold restart it takes 15-30 minutes to get DNS/Internet access via ethernet-cable and/or Wifi.
It did not take more than 30-60 seconds with my previous router.
Where I live is fiber-connection and I live in an apartment, so the switch is not in my own apartment. It´s just an etherhet-socket in the wall.
I understand it could be the switch in the building that is the cause of the problem, but why would my previous router take 30-60 seconds to get internet-access after a cold restart then?
The hAP ax³ AX has the latest RouterOS 7.23.2. (The problem has been present on several RouterOS-versions).
Problem:
It takes too long to get interent access.
My question:
Is there any way I can speed up the time it takes to get internet access?
I have tried Ipconfig-commands like refresh DNS.
Dear all, thank you all for your kind replies.
I will test the suggestions I got in replies.
However I can´t test it until in a few days. I will reply back here of the results after the tests.
It is possible I have induced the problem myself by simply cutting the power, without first turning off the router with the power-switch. Or maybe the particular switch in the buildning I live don´t like my router because of something.
As mentioned, the router is working great otherwise! I also like the support of updates in the RouterOS-versions, so we can have a long-lasting security- and improvements-updates. I also like to support MikroTik as a company.
With best regards
If anyone else have any input I greatly appreciate it.
With best regards
Hi! Thank you for your kind reply.
I did not think about checking the logs, that´s a great idea.
In the logs, the first post is "Buffer: memory, Topics: system - error - critical, Message: router rebooted without proper shutdown, probably power outtage.
No message about DNS.
Regarding WAN there is severel occasions with "ether1 link up (speed 100M, full duplex)", then just 1 second later "ether1 link down"
Ca 21 minutes after power on / first log-entry there is a text "client1 on ether1 got IP address xxx.xx.xxx.xxx in topic "dhcp".
My main hypothesises is that the switch in the building I live either don´t like this particular router, or I am doing a wrong action by simply cutting the power when I turn of the router, instead of doing a shutdown via the button on the router or via Winbox. However all my other routers have never had any problem with just power outage.
I will, however, try to shut down via the power-button and see if it helps.
Hi! I don´t get any internet on Wifi or via Ethernet cable, on my phone or computer. I am able to connect the phone to wifi, but there is not internet. The computer says it´s connected via Ethernet, but not internet.
Hi! Thank you for your kind reply.
Yea, I think it´s also not normal.
I have standard configuration, I haven´t changed anything and this problem was present since the first time I started the router. But only after power loss, meaning not turning off via button or via Winbox.
I don´t know what information in the config-file is sensitive except passwords, so I don´t feel comfortable sharing it online. I am quite new on router-configs so I don´t know much about many of the config-settings.
Check other log entries about ether1 ... what is delay before next "ether1 link up" and what is the speed and duplex mentioned then?
It could be that the switch doesn't like the 2.5Gbps rate which your hAP ax3 supports on ether1 port while your previous router might not have? Or the switch actually likes it but the apartment building UTP installation doesn't conform to requirements to have 2.5Gbps reliably running and both sides (switch and hAP ax3) then struggle with autonegotiaon of reliably working connection.
You can try to remove 2.5Gbps from advertised speeds on ether1 and see if that helps in any way.
Dear all. Thank you for your help. It is now solved with a new Cat 6 cable to WAN and I get internet connection within seconds after a cold restart.
My old cable (cat 5e) seems to have been defective, or not having a solid mechanical connection to the ports.
I saw in logs that the Ether1 connection (WAN) went up and down several times, and negotiated 10mbit- och 100mbit randomly. Pinging 1.1.1.1 as well as the Switch IP-adresses resulted tin 65-80% package lost.'
The computer actually said there WAS internet connection, but could not reach any homepages
- I actually asked ai for help, and it suggested possible defect cable. I then took out the old cat 5e WAN-cable and inserted it again and made a cold restart. This time it worked fine.
I bought a new Cat 6 cable and have changed to this cable now. After cold reboot it works within seconds. So I believe it was the old Cat 5e-cable that had become defective, somehow.
Just not according to the reference I linked (which may well be wrong, or ignored 2.5G outright) . . . (I have to admit, though, that I didn't do an exhaustive search, since I consider 2.5 largely useless . . . ).
Thank you for clarifying this. I also asked ai about that when troubleshooting and according to ai Cat 5e should have been good enough for my 100 mbit down and 10 mbit up internetspeed.
I need to rehearse and study ethernet and newer IP Cat specs However: it was not easier when I first tried to set up LAN in the old good times with coaxial cable, BNC T-connectors and terminators. We never had all computers found, I believe, at most 75%, during these LAN:s. It was weekend LAN, playing multiplayer-games like FPS or strategy games.
For the old cat 5e-cable I did try to disable the 2.5G and only have 1G on Ether1 (WAN), in Winbox-settings. It still had the same problem with 65-80% packet loss at ping and dropping and renegotiating several times between 10 and 100 mbit. It is perhaps possible that it still was the cat 5e-rating that was the problem.
However, also the new cat 6-cable only get 100 mbit negotiated with the switch. I don´t know if it´s because my internet-speed is 100 mbit up and 10 down, or if the buildings switch only has 100 mbit-ports. I imagine the switch has 1G-ports, because it is possible to order internetspeed up to 1000 mbit down and 1000 mbit up.
This. When devices, connected using cat 5e cable, only autonegotiate 10Mbps or 100Mbps, it means that either cable is generally toast (and is uncapable of passing higher frequencies required for 1Gbps operations) or that one of the two extra pairs (1Gbps and above use all 4 pairs, 10/100Mbps use only two pairs) is not connected. And the symptom seen by @Sirius (high packet loss even with 10/100 Mbps negotiated) strongly points at the issue #1 (cable being generally toast). Slightly misleading information by @Sirius was that setup was working fine when using his previous router.
Hi! Thank you for your reply. I agree.
I also agree that it was misleading information, if the cable was bad with the old router. But. It was the truth, as I remember it. It did work fine. Only after I changed router did it get this bad with ca 30+ minutes to be able to access homepages.
Maybe the cable was bad a long time and the disconnection of the cable from the old router made the cable worse... or maybe the cable simply had better physical connection with the old router. I don´t know.
I do know that it works now and I have learned something valuable. Don´t assume that cables are ok, even if they have been "fine" before and seemed to work with previous equipment. It´s also an easy action to change to a new cable in troubleshooting - I should have done that earlier, after seeing the negotiation-problems in the logs.
this can happen (not very often, but it does) ... every plug and port is manufactured up to a spec ... but it can happen that if both entities are actually on the opposite border of spec, there can be interoperability issue.
Another issue (AFAIK shouldn't be with hAP ax3 but one never knows) can be that some device uses lower Tx power (e.g. due to EEE) and that causes too low signal levels on Rx side. In such case using a better cable can help.