Hello,
in all ROS versions i tried the CPU is cycling to 100%, tried on all RB(112.532.333.133), everywhere same / from time to time CPU go to 100%.
The very bad situation is on RB112 where when used ROS 2.9.38 i got 18% CPU load, when used 3.0 ht ecpu is almost all the time on 100% / when connected with winbox and i cant do almost anything . Disabled all packages except system, routerboard, wireless and security.
ANy way i think there must be some kind of bug as the CPU is ticking on all platforms to 100%, when used on 112 it is far slower than 2.9.38.
Any idea ?
Do you have any performance change? Or you just jump out of chair because one number that is background service and don’t have to give precise per second result???
Average CPU load per 5min - thats where the usable information is…
Same problem 3rc13. RB600 no wirelles cards no setting and 100 percent CPU.
With settings same problem and + disconnecting any clients connected → disconnected ->reasociating.
hello all i saw that ROS is now 3.0 so i upgrade to 3.0 hope that thing will get better and i could use l7 at last same configuration on my P4 system with 1G RAM and the CPU usage rise to 100% i downgrade and 1% and some times rise to 15% what could be the problem ? is it a BUG?
MT please to do some thing.
Sure thats an issue with 3.0 , also disconnectiing clients. See i reported it long time ago, got just a lame answer i swear about cpu load.
Indeed you see a lot of channels missing in country settings - i think this is a policy for people to buy superchannel licence.
I remember times when i bought WAR boards from Valemount networks. The boards were completely unusable, locking after few hours itsel, making a lot of troubles and when i tried to solve the problem i was getting answers like this, sometime i got answer about wrong power supply etc . I switched to Mikrotik and RB’s.
Upgrade will not solve this neither firmware upgrade - for me there is some kind of bug in kernel code + unstable wireless drivers. I think when you are not pushing the hardware to higher speeds, this bugs are not present, but try to run it in live environment, over 10 Mbits of real traffic on wireless interface and you will see after few hours. But for lower speeds i dont need to use new RB’s .
Sorry but, its better that a lot of people will report this bug , then maybe it will be solved. But then a lot of people will be unhappy and trying to “solve” the problem by config but i dont think this is a way for product which i paid - also the config will NOT solve the problem completely. The v3 is in early beta - not stable and can cause lot of people to think about stable systems like Canopy.
There is no help, the need is to fix the bugs. I cannot afford to test new SW on my customers and hope this version will finally work.
I was very satisfied with MK until 3,0 comes and same and same bugs repeated in RC’s .
is this 100% cpu causing problems, or is it simply a display issue ? My guess is that its just displaying improperly. I also notice the same thing sometimes, 1-2% CPU and then 100% cpu, and then back. Clearing the winbox saved cache seems to help a little - and truthfully I think it’s winbox. SSH in and see if its the same problem.
It’s not just a display issue, I ran into same bug on RB112 with ROS 3 betas. Took forever to be able to get in and downgrade to get things stable again. From further testing, the NTP package was responsible, and partially so was auto-supout (something (NTP?) was crashing and causing supout to be written constantly which is very CPU intensive). It’s not good news to see the same bug reported in the final version too .
I dont think so it is NTP, in newer 3,0 NTP is only simple. But
try to enable memory meter in winbox and log in into - you will see something is eating memory fast , and when dropped under around 3M then the RB is getting very very slow.
Just to repeat - 2,9 is working ok, so posts like uninstall everything etc are for me unusable.
Im happy for release “stable” 3,0 so more people will report the bugs and maybe they will be fixed faster .
.