Asus RT-AC68U 386.2_6 low on free nvram

[AppleBag]

I'm having it myself:

62860 / 65536 should be fine, shouldn't it?

size: 62885 bytes (2651 left)
2020 custom_clientlist
931 nc_setting_conf
734 wl0_maclist_x
733 wl_maclist_x
562 dhcp_staticlist
549 rc_support
415 dhcp_hostnames
410 wl1_maclist_x
376 vts_rulelist
237 webdav_smb_pc
171 dnsfilter_rulelist
166 subnet_rulelist
151 wollist
131 wl1_chansps
120 qos_rulelist
117 pptpd_clientlist
112 vlan_rulelist
92 1:pa5ga2
92 1:pa5ga1
92 1:pa5ga0

I've already gone through and edited custom_clientlist down to bare minimum (removed all spaces, made labels much shorter and thus more cryptic, etc.) and still get the flashing exclamation notice about low nvram.

 

[node]

I too running the latest RT-AC68U Merlin firmware and getting the low NVRAM exclamation flashing. Check indeed low.
My issue is I am 4000 mile from my router and cannot do a factory reset and login back in to reconfigure. Attached is my NVRAM variables, can I wipe or move the heavy users out?

 

[ColinTaylor]

I too running the latest RT-AC68U Merlin firmware and getting the low NVRAM exclamation flashing. Check indeed low.
My issue is I am 4000 mile from my router and cannot do a factory reset and login back in to reconfigure. Attached is my NVRAM variables, can I wipe or move the heavy users out?

How low is "low". nvram usage doesn't change much once the router has been configured and is up and running. So unless your nvram usage is >99% it's probably safe to leave it. At least until a time when you're on site.

That said, it should be safe enough to clear down that largest nvram variable.

nvram set MULTIFILTER_TMP_T=""

 

[node]

Thanks Colin, was a little more than 63K from the 65.2K, ran the code, now sitting at 52.2K. A significant gain. Thanks a mill.

node

 

[rivera]

Is there any simple way to get a remaining size of NVRAM?
Surely I may use this way:

nvram show | awk '{print length(), $0 | "sort -n -r"}' | cut -d"=" -f 1 | head -n 20

But I am only interested in size, not other details...
Running this command takes some time, I need something more fast & simple.

 

[ColinTaylor]

Is there any simple way to get a remaining size of NVRAM?
Surely I may use this way:

nvram show | awk '{print length(), $0 | "sort -n -r"}' | cut -d"=" -f 1 | head -n 20

But I am only interested in size, not other details...
Running this command takes some time, I need something more fast & simple.

nvram show | grep ^size

 

[glenw]

Is there any simple way to get a remaining size of NVRAM?
Surely I may use this way:

nvram show | awk '{print length(), $0 | "sort -n -r"}' | cut -d"=" -f 1 | head -n 20

But I am only interested in size, not other details...
Running this command takes some time, I need something more fast & simple.

Hi,
You could use a monitoring app to keep an eye on free memory. I use a simple little one called Algorius but have used PRTG in the past to do the same sort of thing. I just monitor a few devices and get email alerts if anything strange happens. You can choose what to monitor - only RAM if that's all you care about.

 

[rivera]

You could use a monitoring app to keep an eye on free memory

I am using Home Assistant and getting some metrics from routers via SNMP.
Only getting NVRAM size caused some problems since my previous command took some time for execution.
Now it works much faster!

 

[martinr]

Take a look at YazDHCP as per post #2. That may do it, I'm not sure.

Just worked for me, thanks Colin, after I got the message post update to 386.5 (from 386.2). Anything to avoid a factory reset.

 

[Woody1]

Is there a way to move sshd_authkeys to jffs? This is eating up a lot of nvram for me.

 

[unsynaps]

Install YazDHCP from within AMTM and it moves your reserved IP lan information to /jffs usually helps people that have lots of reserved IP's

Or just do your reservations in dnsmasq.conf.add

dhcp-host=00:1F:29:2B:19:C1,192.168.50.18,printer

 

[Woody1]

Or just do your reservations in dnsmasq.conf.add

dhcp-host=00:1F:29:2B:19:C1,192.168.50.18,printer

Does that help with sshd-authkeys?

 

[RMerlin]

Is there a way to move sshd_authkeys to jffs? This is eating up a lot of nvram for me.

sshd_authkeys should already be on JFFS. Check if /jffs/nvram/sshd_authkeys exists, if it does then you can unset the nvram variable.

If it isn't there, then it means that it wasn't moved for that model. Nothing you can do about it.

 

[Woody1]

sshd_authkeys should already be on JFFS. Check if /jffs/nvram/sshd_authkeys exists, if it does then you can unset the nvram variable.

If it isn't there, then it means that it wasn't moved for that model. Nothing you can do about it.

I have an RT-AC68P. /jffs/nvram exists, but there are no files or folders there.

 

[RMerlin]

I have an RT-AC68P. /jffs/nvram exists, but there are no files or folders there.

Then the move to JFFS for that setting must have been only on newer models.

 

[lokester]

Regarding moving the reserved IP's to /jffs/configs/dnsmasq.conf.add, is there still a limit of 64 maximum entrys ?

 

[ColinTaylor]

Regarding moving the reserved IP's to /jffs/configs/dnsmasq.conf.add, is there still a limit of 64 maximum entrys ?

No.

 

[jmsfll]

How low is "low". nvram usage doesn't change much once the router has been configured and is up and running. So unless your nvram usage is >99% it's probably safe to leave it. At least until a time when you're on site.

That said, it should be safe enough to clear down that largest nvram variable.

nvram set MULTIFILTER_TMP_T=""

Hi.... Hoping you or someone can help me out. I just updated to Merlin and getting the warning. Ran the script RMerlin posted and reduced it a little but still over 63k
Also ran the command which I think is outputting nvram variables:
In my case is it safe to delete anything or any other suggestions? Thanks in advance!

admin@RT-AC68P-9878:/jffs/scripts# nvram show | awk '{print length(), $0 | "sort -n -r"}' | cut -d"=" -f 1 | head -n 20
size: 63766 bytes (1770 left)
931 nc_setting_conf
573 rc_support
312 custom_clientlist
164 subnet_rulelist
131 wl1_chansps
129 wps_env_buf
124 vts_rulelist
120 qos_rulelist
114 dhcp_staticlist
112 vlan_rulelist
92 1:pa5ga2
92 1:pa5ga1
92 1:pa5ga0
89 vpn_server2_ncp_ciphers
89 vpn_server1_ncp_ciphers
89 vpn_client5_ncp_ciphers
89 vpn_client4_ncp_ciphers
89 vpn_client3_ncp_ciphers
89 vpn_client2_ncp_ciphers
89 vpn_client1_ncp_ciphers

 

[PeterR]

Hi.... Hoping you or someone can help me out. I just updated to Merlin and getting the warning. Ran the script RMerlin posted and reduced it a little but still over 63k
Also ran the command which I think is outputting nvram variables:
In my case is it safe to delete anything or any other suggestions? Thanks in advance!

admin@RT-AC68P-9878:/jffs/scripts# nvram show | awk '{print length(), $0 | "sort -n -r"}' | cut -d"=" -f 1 | head -n 20
size: 63766 bytes (1770 left)
931 nc_setting_conf
573 rc_support
312 custom_clientlist
164 subnet_rulelist
131 wl1_chansps
129 wps_env_buf
124 vts_rulelist
120 qos_rulelist
114 dhcp_staticlist
112 vlan_rulelist
92 1:pa5ga2
92 1:pa5ga1
92 1:pa5ga0
89 vpn_server2_ncp_ciphers
89 vpn_server1_ncp_ciphers
89 vpn_client5_ncp_ciphers
89 vpn_client4_ncp_ciphers
89 vpn_client3_ncp_ciphers
89 vpn_client2_ncp_ciphers
89 vpn_client1_ncp_ciphers

[ALPHA] 386.9_alpha1 for WiFi 5 (802.11ac) Models (GT-AC2900, RT-AC68U, and RT-AC86U)

Per Rmerlin in the initial 388.1 Beta thread post. https://www.snbforums.com/threads/asuswrt-merlin-388-1-beta-is-available-for-select-models.81644/ Edit to add: There is also mention of it in the firmware changelog file as well both in the 386.9 (xx-xxx-xxxx) section. And in the 388.1...

www.snbforums.com

 

[e_lasman]

I am wondering what is that "nc_settings_conf" anyway? It is the largest consumer of my NVRAM. I did check the code https://github.com/RMerl/asuswrt-merlin.ng/blob/master/release/src/router/rc/services.c but can't understand what notify center this code is about. Anyone can shed more light on this? Is it safe to unset the value of "nc_settings_conf"?