AX88U high RAM usage, maybe?

[archiel]

I'm now 11 days in on this firmware and free memory is still at 500mb's which was never the case with any previous version so just an fyi

View attachment 38551

Same observation here, on 386.3 (All versions) I would see the memory steadily climb over a few days, then reset down while the swap file gradually increased. Since moving to 386.4 (and full reset and manual rebuild) the free memory is noticeable greater than before and the swap is still nil. FWIW temperatures are marginally higher as is CPU load even when the router is idle.

 

[Col8eral]

RT-AX88U with 3004.388.4

Total ram used 58%

If I plug a drive Into the router over time ram usage goes to 97% . The router slows down, internet speed tests show half the speed I should get.

Unmount the drive and all is well.

Re mount the drive and the process starts again.

Is this fixable or can't these routers handle attached storage.

Many thanks

 

[ColinTaylor]

RT-AX88U with 3004.388.4

Total ram used 58%

If I plug a drive Into the router over time ram usage goes to 97% . The router slows down, internet speed tests show half the speed I should get.

Unmount the drive and all is well.

Re mount the drive and the process starts again.

Is this fixable or can't these routers handle attached storage.

Many thanks

You need to determine what process is accessing the USB drive when you plug it in. It could be the Media Server in which case the activity will stop after it's finished indexing the contents of the drive.

 

[Col8eral]

It seems to go to 97% and then just hangs there. I've waited but it doesn't change..then the performance of the router drops off and services slow down.

 

[ColinTaylor]

It seems to go to 97% and then just hangs there. I've waited but it doesn't change..then the performance of the router drops off and services slow down.

The 97% RAM use is exactly normal for that router when reading or writing more than ~600MB of data from the drive as it's being cached in memory. The problem with the router slowing down or hanging is more likely to be the CPU load associated with the IO activity.

SSH into the router and run top. That might give you a clue as to what process is causing the problem.

 

[Col8eral]

Thanks for that. It seems if I upload a fe or download a file or view an image or two that's when the ram fills up. Even if I stop accessing the drive, the router performance just stays bad.

What is top and how do I run it if you don't mind me asking. Cheers.

 

[ColinTaylor]

Thanks for that. It seems if I upload a fe or download a file or view an image or two that's when the ram fills up.

Yes that's to be expected and not a problem in itself.

Even if I stop accessing the drive, the router performance just stays bad.

What is top and how do I run it if you don't mind me asking. Cheers.

top is a Linux command that shows you the "top" processes. You need to enable SSH in the router's GUI and then log into the router, either with a client like PuTTY or from the Windows/Linux command line.

ssh admin@192.168.1.1

Then run the top command:

 

[ColinTaylor]

Also check the router's System Log. In particular look for USB related error messages.

 

[Ripshod]

I've attached a screenshot from scMerlin that demonstrates my actual RAM usage (specs as my signature). I only use my ssd for entware and scripts and the actual RAM used is only that high because I use Diversion and dnsmasq logging.
Because I don't access the drive other than SFTP for a few logs (USB sharing by SMB and FTP is disabled) my cache is relatively small, but with file sharing enabled and larger transfers that would easily expand the cache to use nearly all the spare space.
The beauty of cache is if the system needs more RAM the cache will be reduced to expand the available space for the system with no noticeable effect on the router or the user experience. There is no way this operation would cause any slowness unless the usb drive has errors or is just plain failing.
As the saying goes, "In Linux empty ram is wasted ram".
Isn't it time for a sticky on this? This is a perpetual question........

 

[Col8eral]

Here's my results from the top command.

Mem: 866112K used, 37408K free, 5568K shrd, 16K buff, 413596K cached

CPU: 0.5% usr 0.6% sys 0.0% nic 98.7% idle 0.0% io 0.0% irq 0.0% si

Load average: 1.91 1.92 1.97 2/198 4

02rq


The is now ruining slowly and connected devices laggy. Remove the drive and all is well

 

[JGrana]

The top results look reasonable. On my lightly loaded AX86U (with an attached USD SSD drive) I get similar results. CPU usr and sys are low, lots of low priority processes running (nic %). My load average is a bit lower (around 1.5).

If you have Entware installed, you could try iostat to see if there are lots of reads/writes to your drive.
Something like:

$ iostat -d 1 30

That would show only device activity (USB drive) once a second for 30 seconds.

iostat is part of the opkg package sysstat:

$ opkg install sysstat

 

[ColinTaylor]

Here's my results from the top command.

Next time can you post a picture of the entire window rather than just the first couple of lines. Also check the System Log.

 

[Col8eral]

Here's all of the output from TOP command

Mem: 871992K used, 31528K free, 3884K shrd, 16K buff, 437680K cached
CPU: 1.4% usr 1.4% sys 0.0% nic 97.0% idle 0.0% io 0.0% irq 0.0% sirq
Load average: 2.09 2.14 2.04 1/196 24970
PID PPID USER STAT VSZ %VSZ CPU %CPU COMMAND
24927 24910 admin R 3424 0.3 3 1.4 top
1305 1 admin S 20056 2.2 1 0.0 amas_lib
1933 1 admin S 19408 2.1 3 0.0 wred -B
1254 1 admin S 18960 2.1 3 0.0 conn_diag
321 1 admin S 18524 2.0 1 0.0 /bin/swmdk
1923 1 admin S < 18160 2.0 1 0.0 dcd -i 3600 -p 43200 -b -d /tmp/bwdpi/
1252 1 admin S 17820 1.9 0 0.0 roamast
1282 1254 admin S 17804 1.9 2 0.0 amas_portstatus
2430 1 admin S 16592 1.8 3 0.0 minidlna -f /etc/minidlna.conf -r
1286 1 admin S 15520 1.7 3 0.0 cfg_server
1083 1 admin S 14944 1.6 3 0.0 nt_center
1072 1 admin S 14732 1.6 3 0.0 /sbin/netool
1208 1 admin S 14064 1.5 1 0.0 networkmap --bootwait
1070 1 admin S 13892 1.5 0 0.0 nt_monitor
1 0 admin S 13780 1.5 2 0.0 /sbin/init
1172 1 admin S 12684 1.4 2 0.0 watchdog
1176 1 admin S 12684 1.4 1 0.0 sw_devled
1174 1 admin S 12684 1.4 1 0.0 alt_watchdog
1051 1 admin S 12684 1.4 0 0.0 /sbin/wanduck
1256 1 admin S 12684 1.4 0 0.0 erp_monitor
1395 1 admin S 12684 1.4 3 0.0 usbled
1980 1 admin S 12684 1.4 3 0.0 bwdpi_wred_alive
1214 1 admin S 12684 1.4 1 0.0 pctime


And here's some of the system logs:

Oct 4 16:08:53 wlceventd: wlceventd_proc_event(695): eth7: ReAssoc 10:63:C8:53:FA:4B, status: Successful (0), rssi:-69
Oct 4 16:08:55 wlceventd: wlceventd_proc_event(645): eth7: Deauth_ind 84:FD1:E7:5C:FC, status: 0, reason: Station requesting (re)association is not authenticated with responding station (9), rssi:0
Oct 4 16:08:55 wlceventd: wlceventd_proc_event(662): eth7: Disassoc 84:FD1:E7:5C:FC, status: 0, reason: Disassociated because sending station is leaving (or has left) BSS (8), rssi:0
Oct 4 16:08:57 wlceventd: wlceventd_proc_event(685): eth7: Auth 84:FD1:E7:5C:FC, status: Successful (0), rssi:0
Oct 4 16:08:57 wlceventd: wlceventd_proc_event(695): eth7: ReAssoc 84:FD1:E7:5C:FC, status: Successful (0), rssi:-77
Oct 4 16:09:39 wlceventd: wlceventd_proc_event(662): eth7: Disassoc 84:FD1:E7:5C:FC, status: 0, reason: Disassociated because sending station is leaving (or has left) BSS (8), rssi:0
Oct 4 16:10:11 wlceventd: wlceventd_proc_event(645): eth7: Deauth_ind 10:63:C8:53:FA:4B, status: 0, reason: Unspecified reason (1), rssi:0
Oct 4 16:10:11 wlceventd: wlceventd_proc_event(685): eth7: Auth 10:63:C8:53:FA:4B, status: Successful (0), rssi:0
Oct 4 16:10:11 wlceventd: wlceventd_proc_event(695): eth7: ReAssoc 10:63:C8:53:FA:4B, status: Successful (0), rssi:-69
Oct 4 16:13:18 wlceventd: wlceventd_proc_event(645): eth7: Deauth_ind 10:63:C8:53:FA:4B, status: 0, reason: Unspecified reason (1), rssi:0
Oct 4 16:13:18 wlceventd: wlceventd_proc_event(685): eth7: Auth 10:63:C8:53:FA:4B, status: Successful (0), rssi:0
Oct 4 16:13:18 wlceventd: wlceventd_proc_event(695): eth7: ReAssoc 10:63:C8:53:FA:4B, status: Successful (0), rssi:-69
Oct 4 16:16:26 wlceventd: wlceventd_proc_event(645): eth7: Deauth_ind 10:63:C8:53:FA:4B, status: 0, reason: Unspecified reason (1), rssi:0
Oct 4 16:16:26 wlceventd: wlceventd_proc_event(685): eth7: Auth 10:63:C8:53:FA:4B, status: Successful (0), rssi:0
Oct 4 16:16:26 wlceventd: wlceventd_proc_event(695): eth7: ReAssoc 10:63:C8:53:FA:4B, status: Successful (0), rssi:-69
Oct 4 16:19:33 wlceventd: wlceventd_proc_event(645): eth7: Deauth_ind 10:63:C8:53:FA:4B, status: 0, reason: Unspecified reason (1), rssi:0
Oct 4 16:19:34 wlceventd: wlceventd_proc_event(685): eth7: Auth 10:63:C8:53:FA:4B, status: Successful (0), rssi:0
Oct 4 16:19:34 wlceventd: wlceventd_proc_event(695): eth7: ReAssoc 10:63:C8:53:FA:4B, status: Successful (0), rssi:-69
Oct 4 16:22:41 wlceventd: wlceventd_proc_event(645): eth7: Deauth_ind 10:63:C8:53:FA:4B, status: 0, reason: Unspecified reason (1), rssi:0
Oct 4 16:22:41 wlceventd: wlceventd_proc_event(685): eth7: Auth 10:63:C8:53:FA:4B, status: Successful (0), rssi:0
Oct 4 16:22:41 wlceventd: wlceventd_proc_event(695): eth7: ReAssoc 10:63:C8:53:FA:4B, status: Successful (0), rssi:-68
Oct 4 16:22:52 wlceventd: wlceventd_proc_event(645): eth7: Deauth_ind 10:63:C8:53:FA:4B, status: 0, reason: Unspecified reason (1), rssi:0
Oct 4 16:22:52 wlceventd: wlceventd_proc_event(685): eth7: Auth 10:63:C8:53:FA:4B, status: Successful (0), rssi:0
Oct 4 16:22:52 wlceventd: wlceventd_proc_event(695): eth7: ReAssoc 10:63:C8:53:FA:4B, status: Successful (0), rssi:-67
Oct 4 16:25:39 wlceventd: wlceventd_proc_event(685): eth7: Auth 84:FD1:E7:5C:FC, status: Successful (0), rssi:0
Oct 4 16:25:39 wlceventd: wlceventd_proc_event(645): eth7: Deauth_ind 84:FD1:E7:5C:FC, status: 0, reason: Unspecified reason (1), rssi:0
Oct 4 16:25:39 wlceventd: wlceventd_proc_event(685): eth7: Auth 84:FD1:E7:5C:FC, status: Successful (0), rssi:0
Oct 4 16:25:39 wlceventd: wlceventd_proc_event(722): eth7: Assoc 84:FD1:E7:5C:FC, status: Successful (0), rssi:-69
Oct 4 16:26:00 wlceventd: wlceventd_proc_event(645): eth7: Deauth_ind 10:63:C8:53:FA:4B, status: 0, reason: Unspecified reason (1), rssi:0
Oct 4 16:26:00 wlceventd: wlceventd_proc_event(685): eth7: Auth 10:63:C8:53:FA:4B, status: Successful (0), rssi:0
Oct 4 16:26:00 wlceventd: wlceventd_proc_event(695): eth7: ReAssoc 10:63:C8:53:FA:4B, status: Successful (0), rssi:-68
Oct 4 16:27:52 wlceventd: wlceventd_proc_event(645): eth7: Deauth_ind 10:63:C8:53:FA:4B, status: 0, reason: Unspecified reason (1), rssi:0
Oct 4 16:27:52 wlceventd: wlceventd_proc_event(685): eth7: Auth 10:63:C8:53:FA:4B, status: Successful (0), rssi:0
Oct 4 16:27:52 wlceventd: wlceventd_proc_event(695): eth7: ReAssoc 10:63:C8:53:FA:4B, status: Successful (0), rssi:-67
Oct 4 16:31:00 wlceventd: wlceventd_proc_event(645): eth7: Deauth_ind 10:63:C8:53:FA:4B, status: 0, reason: Unspecified reason (1), rssi:0
Oct 4 16:31:00 wlceventd: wlceventd_proc_event(685): eth7: Auth 10:63:C8:53:FA:4B, status: Successful (0), rssi:0
Oct 4 16:31:00 wlceventd: wlceventd_proc_event(695): eth7: ReAssoc 10:63:C8:53:FA:4B, status: Successful (0), rssi:-69
Oct 4 16:34:07 wlceventd: wlceventd_proc_event(645): eth7: Deauth_ind 10:63:C8:53:FA:4B, status: 0, reason: Unspecified reason (1), rssi:0
Oct 4 16:34:07 wlceventd: wlceventd_proc_event(685): eth7: Auth 10:63:C8:53:FA:4B, status: Successful (0), rssi:0
Oct 4 16:34:07 wlceventd: wlceventd_proc_event(695): eth7: ReAssoc 10:63:C8:53:FA:4B, status: Successful (0), rssi:-69
Oct 4 16:36:11 dropbear[24895]: Password auth succeeded for 'admin' from 192.168.1.12:60967
Oct 4 16:37:15 wlceventd: wlceventd_proc_event(645): eth7: Deauth_ind 10:

 

[ColinTaylor]

That all looks normal. But I suggest you go to USB Application > Servers Center > Media Server and turn off the Media Server.

 

[Col8eral]

Many thanks. Turned off Media Server. Reconnected USB drive. RAM jumps from 55% to 66%. Play a video from the USB drive RAM jumps from 66% to 97% and stays there. Then everything starts slowing down again until I unplug the drive, then all back to normal

 

[Col8eral]

So, it looks like if I want to use a connected USB drive, every time I access files on the drive or upload or download files, I have to clear the cache manually. Either of these two seem to work:

sync && echo 3 > /proc/sys/vm/drop_caches
or
free && sync && echo 3 > /proc/sys/vm/drop_caches && free

A bit tedious to have to do this each time. I have been using the rive without any RAM impact for about a year. Perhaps the FW update had an impact. That's the only thing that's just during that time frame

 

[Ripshod]

Many thanks. Turned off Media Server. Reconnected USB drive. RAM jumps from 55% to 66%. Play a video from the USB drive RAM jumps from 66% to 97% and stays there. Then everything starts slowing down again until I unplug the drive, then all back to normal

How are you playing videos from the attached drive with the media server disabled?

 

[Col8eral]

How are you playing videos from the attached drive with the media server disabled?

Hi. Either playing from Windows file manager using windows media player. Or via Plex. The end result in terms of ram if I just upload or download files.

 

[Ripshod]

Hi. Either playing from Windows file manager using windows media player. Or via Plex. The end result in terms of ram if I just upload or download files.

So you're still filling the cache, just via a different method.
Having the same fimware and hardware I'm failing to understand why your cache isn't released as ram is needed. What scripts do you have installed (if any), official or unofficial?

 

[Col8eral]

If the drive isn't plugged in the ram sits at 55% so I I don't think the scripts are doing it. It's been fine for a year or so. The only change was the recent fw update.

I run a script to monitor CPU temp and another script to send me an email when a VPN client connects.

Everything else is stock.