I have serveral RB133, and until yesterday all of them were working fine, now, one is giving me problems with cpu, it was constantly on 100% cpu usage, i didnt changed the config, ran torch and the traffic that used to passed there was the same as should, no torrents no attacks nothing, this routerboard is on a top of a building that i only have access 1 or 2 times a year… so its a litle crappy getting there to do some site maintence :S
Im using the 3.18 ( 3.20 gave me some problem on the signal quality by dramatically decreasing the signal quality and tx/rx of my clients ).
It has one EMP-3602 and one winstron cm9 in ap bridge mode
Its main purpose is only to be a wireless apbridge without any routes configured our firewall rules.
Ive downgraded and upgraded serveral times without getting any progress on this problem, ive even tried to load the las backup, but it was the same as talking to a wall
Hi,I note that sometimes problems with PCI signaling is reason for RB unstable work and 100% CPU load. Try disable wireless card one by one and look what’s happen. Else you can clean mPCI card gold fingers with something like acetone .
Ive manached to get a sneek on the condition its 100% ok ( lol as same as cpu usage ), no dust no loose parts, ive mesured the power supply and the power cable that came form the build and it was ok, ive recleaned up all the contcts with a cleanup spray for electrical contacts from philips ( cant remember de code of it, but is the same that i use to clean up pc´s boards ,memorys and expansion cards ), ive even disabled the cards one by one as you suggested, the results are the same… im getting a litle bit desperate with this… my clients are already biting my legs :S
Were i can send the support.rif so someone could help me try to figure this out ?
Ive managed to drecrese de cpu usage ! XD , now its 13/14%, ive done a check-installation and a check-disk, didnt report anything strange, but the funny thing is that appeard to temp. solve this problem ! , im sending the support.rif that ive made after doing this, could be something that isnt right..
The problem came back, the rb 133 simply freezes without anything made on it, cpu 100%…
Ive sent the support.rif file ,and i had a quick reply ( awsome ), saw the suggestion that support gave it to me to see if it was from the watchdog but i get the same thing, still trying to find what is wrong with it.. Oh well, we will try until get some good news from it
Hi there, this RB133 manages only a few clients, ( 20 top ), ive got no special services running on it, this board was working 100% fine until that day, with the same amout of clients and had more services running, the cpu was at top 43% when performing a supout.rif .
The hardware is fine, all cards are working 100% no drops no probs..
I have a second RB133 doing exactly the same has that one above, and the cpu is only between 7% ( normal load ) - 45% ( preforming a supout.rif ).
So this excessive load doesn’t make any sense… im still waiting for a valid solution by mikrotik support team ( send out my supout.rif ), if there isn’t one, i have to try netinstall and reconfigure everything from scratch..
The only problem is that node is on a top of a tall building which i wont have the key soon..
( Sorry about my crappy English )
PS: below you can see the load of the second RB133 with the same cfg and the same number of clients connected
Ive noticed that removing the snmp made quite some improvements to me too, in fact, since i disabled it ( ± 14 hours ago ), the RB133 is working as should be as you can see on the img. below.
