I 've just encountered the Flash CPU usage problem.
Flash 80% CPU, still responding to pings, ssh/webgui not responding, wifi clients can’t connect, can still login via winbox, very slow, able to reboot. Happened this morning, not sure what time but I only realized it at around 11am.
Overnight was downloading just one well-seeded torrent over lan and seeding about 10 others but less than 1mbps upload in total. Nothing else. No Skype calls or anything like that.
Doesn’t look good for this router. Will be changing back to my RB750GL + Airport Extreme once I’m done experimenting with it.
It’s kinda happening again. For the last 5 mins or so, I was watching the Flash process in Profiler fluctuating between 20 & 80%. Now it seems to have stopped. No entries were written to the log. Sector writes in Resources did not increase. Nothing on the network stressing the router. No downloads going on.
I’m also seeing this problem with my RB751U-2HnD. 100% flash usage in profiler, with random reboots/inability to login with Winbox. I’ve tried disabling SNMP, removing all graphing, uninstalling ntp/user-manager/ipv6/routing and ensuring that transparent proxy is not used for hotspot.
This RB acts as a hotspot for a coffee shop, with a max of about 15 users who use only the ‘trial’ functionality. Also has a VAP on same wlan card, but doesn’t see much traffic.
Only occurrs when a few users are logged on and actually using the hotspot. Not using the dude or anything else that polls the router.
I had this problem as well with my RB751U-2HnD, but it seemed to go away around ROS v 5.16-5.17. Exact same symptoms as everyone else has reported above. 80% or so CPU usage in the “flash” category with no hotspot, usermanager, proxy, ntp, or anything else mentioned to check above. I thought I was home free when the issue seemed to go away a few months ago. Just the other day I upgraded to 5.20 and the issue with the high CPU in the “flash” category is back. I also did a /system check-disk and did a full format and re-install using netinstall, no help.
I’m seeing the same thing – 100% CPU spikes on my RB751U-2HnD running 5.22, associated with ‘flash’ in the profiler.
There’s a dozen or so clients on the network at a given time, largely doing low-bandwidth stuff through their ISP.
The only thing I use to poll the router is Winbox itself, which I started using when I started seeing “hangage”.
I have remote syslog configured for critical/error/warning, but that’s not much logging. No scripts.
There’s occasionally Apple gear on the network, but mostly iPads/iPhones.
nov/13/2022 08:53:51 system,error,critical router was rebooted without proper shutdown
nov/13/2022 08:53:57 system,error,critical kernel failure in previous boot
nov/13/2022 08:53:57 system,error,critical out of memory condition was detected
Tried clean installation with netinstall, no better results.