What's new

Kamoj Kamoj Add-on Beta testing

  • SNBForums Code of Conduct

    SNBForums is a community for everyone, no matter what their level of experience.

    Please be tolerant and patient of others, especially newcomers. We are all here to share and learn!

    The rules are simple: Be patient, be nice, be helpful or be gone!

I'm rocking a R7800 with FW:V1.0.2.83.2SF (Voxel). Kamoj Add-on:5.4b26.

Code:
Found the following usable USB devices:
sda5 mounted as /tmp/mnt/sda5 (ext4). Label:optware. Model:SanDisk-Ultra Fit. Size:30.8GB
sdb5 mounted as /tmp/mnt/sdb5 (ext4). Label:storage. Model:SanDisk-Ultra Fit. Size:30.8GB
--------------------------------------------------------------------------------
Creating and installing SSH keys in router and writing tar-backup to USB at:/tmp  /mnt/sda5

View attachment 33112

For the sake of keeping this short I'll provide info from USB1 only.

Code:
CONNECTED_USB_DEVICES='sda sdb'
USB_BUS='2'
USB_CAT='3.20'
USB_DEV='sda'
USB_DEVID='/dev/sda5'
USB_FILESYS='ext4'
USB_FREEMEM='26.792 GB'
USB_IVL='0ms'
USB_LABEL='optware'
USB_MANUF='SanDisk'
USB_MODEL='Ultra Fit'
USB_MOUNTID='sda5'
USB_MOUNTING='/tmp/mnt/sda5'
USB_PARTITION_NO='1'
USB_PORT='USB1'
USB_PROD='Ultra Fit'
USB_PWR='224mA'
USB_SERIAL='0501a204158f77fb3dbb4559a229ef97e6a52764ee1092f344b1ce8dee1ba4078c780000000000000000000088019cc2ff96'
USB_SPEED='5000 Mbps'
USB_TOTMEM='28.289'
USB_TOTMEM_B='29662852.000'
USB_USEDMEM='0.064'
USB_USEDMEM_B='66964.000'
USB_USEDPCT='0%'
USB_UUID='780542cb-9f46-9bb2-5525-3c522cb9740e'
USB_VENDOR='SanDisk'
USB_VOLUME='optware'

The ISSUE is with resetting MOUNT points. Entware script seems to be bound to sda1 and fails for some reason if I run it with one USB flash drive attached (mounted as sda5). Since I have NO slightest clues on how to reset sda5 -> sda1, I'm stuck.

@R.Gerrits really did help me A LOT with information and shell code for manually renaming share names (which cannot be done from GUI), I'm still thankful for that. :)
Ok, thanks for fast response.
If you cant change sda5, you can always add another mounting point sda1 that becomes an alternative to sd5. If you need /tmp/mnt/sda1 of some reason.:
Code:
mkdir -p /tmp/mnt/sda1
mount /dev/sda5 /tmp/mnt/sda1
 
The ISSUE is with resetting MOUNT points. Entware script seems to be bound to sda1 and fails for some reason if I run it with one USB flash drive attached (mounted as sda5). Since I have NO slightest clues on how to reset sda5 -> sda1, I'm stuck.

normally entware doesn't care about the mountpoint.

Firmware should just pass the sda5 as an argument to the post-mount.sh
and then that post-mount.sh should just symlink the /tmp/mnt/sda5/entware/* folders onto /opt/*
(where * is etc, bin, home, root, lib and a few others).

So it this is not working, then I would look what is wrong with /tmp/mnt/sda5/autorun/scripts/post-mount.sh

(or you are using a completely different entware installation method then what Voxel is describing in his readme)
 
@R.Gerrits, I started experimenting with entware_install.sh and my issue was with function format_usb_device(), after unmounting it would not be possible to mount the same flash drive again by the script and further installation files could not be stored - resulting in unrecoverable errors. Unplugging the USB flash drive and inserting it into computer I was left with no file system and had to create EXT4 partition again. So the issue was somewhere along the formatting process. Since it was super late I decided to pre-format flash and comment out the offending function, entware installed fine afterwards.

So, yeah, takeaway is that mounting point doesn't have to be sda1 for it to work.

I will not bother with this for now since eventually Entware was successfully installed and thanks to You I'm able to swap share names from shell within nvram.
 
Hi Kamoj,
I updated to 5.4b26 and found two small issues (both in 1.0.2.83SF and 1.0.2.84SF on R7800):
1. Each time the router is rebooted, the WAN LED turns orange during operation. In order for the LED to return to white color, you have to disconnect and connect the patch cord of WAN port. This does not happen without the add-on installed, nor in the version 5.0 add-on. Regardless of the above speed of ethernet port is still 1GB/s. This does not seem to be a problem with autonegotiation of ethernet ports or with the patchord.
2. I'm connecting to the router interface via https. Then, in the default Adguard Home configuration (without enabled encryption), the Adguard Home link in Kamoj Menu looks like https://routeraddress:8080 , and especially with this port it should be rather http://, or maybe as follows from the current configuration from the AdGuardHome.yaml file

Thanks for all your efforts.
Mark
 
Hi Kamoj,
I updated to 5.4b26 and found two small issues (both in 1.0.2.83SF and 1.0.2.84SF on R7800):
1. Each time the router is rebooted, the WAN LED turns orange during operation. In order for the LED to return to white color, you have to disconnect and connect the patch cord of WAN port. This does not happen without the add-on installed, nor in the version 5.0 add-on. Regardless of the above speed of ethernet port is still 1GB/s. This does not seem to be a problem with autonegotiation of ethernet ports or with the patchord.
2. I'm connecting to the router interface via https. Then, in the default Adguard Home configuration (without enabled encryption), the Adguard Home link in Kamoj Menu looks like https://routeraddress:8080 , and especially with this port it should be rather http://, or maybe as follows from the current configuration from the AdGuardHome.yaml file

Thanks for all your efforts.
Mark
Thank you too!
1. So the WAN port is working even though the WAN LED is orange?
The add-on can set the WAN LED orange in these cases:
OpenVPN Client has failed.
Wireguard Client has failed.
So please go to these submenus and switch off the Client(s) you are not using. You can also check these logs.
If this didn't help:
Do you have any other non-Voxel add-on/software installed, e.g. Entware or Aegis?

2. Correct, it was designed that way and I never had a thought of that it had no secure connection...
My fault. Thank you for reporting! Fixed in next version.
 
Hello Gerrits, tnx for your reply. after few days of investigation i have disabled 11k in Voxel debug session and it seems is sticking now around 70%. i'll keep monitoring for a while.
Hello,

so another update. it seems 11k is not the issue here but adguard v0.106.0-b.1. not sure if it is the beta version but it is filling up the ram
 
Thank you too!
1. So the WAN port is working even though the WAN LED is orange?
The add-on can set the WAN LED orange in these cases:
OpenVPN Client has failed.
Wireguard Client has failed.
So please go to these submenus and switch off the Client(s) you are not using. You can also check these logs.
If this didn't help:
Do you have any other non-Voxel add-on/software installed, e.g. Entware or Aegis?

2. Correct, it was designed that way and I never had a thought of that it had no secure connection...
My fault. Thank you for reporting! Fixed in next version.
Thank you for quick response.
Regarding #1:
Yes. WAN is working when the LED is orange. There are no other software except the add-on and Adguard Home.
Wiregurad Client is not installed (not checked [] Install in General Setting)
OpenVPN Client is not enabled (Red Square in Menu) and not configured, however in accordance to its logs is trying to start, and the last line in log is something like WARNING OpenVPN Client disabled in nvram
Regarding #2:
I think there is worse case in default configuration of Adguard Home after installation. Due to bind 0.0.0.0 you can connect to Adguard Home WEB console using router's WAN port address. You cannot change this from WEB console.
 
Thank you for quick response.
Regarding #1:
Yes. WAN is working when the LED is orange. There are no other software except the add-on and Adguard Home.
Wiregurad Client is not installed (not checked [] Install in General Setting)
OpenVPN Client is not enabled (Red Square in Menu) and not configured, however in accordance to its logs is trying to start, and the last line in log is something like WARNING OpenVPN Client disabled in nvram
Regarding #2:
I think there is worse case in default configuration of Adguard Home after installation. Due to bind 0.0.0.0 you can connect to Adguard Home WEB console using router's WAN port address. You cannot change this from WEB console.
1. So please toggle the OpenVPN Client ON/OFF (Red square) and see what happens. Maybe Netgear's or Voxel's client interferes.
2. Maybe you underestimate the add-on capabilities? ;-)
 
Regarding #2:
I think there is worse case in default configuration of Adguard Home after installation. Due to bind 0.0.0.0 you can connect to Adguard Home WEB console using router's WAN port address. You cannot change this from WEB console.

There is nothing wrong with binding to 0.0.0.0. the fact that you internally can also connect to it via your WAN ip-address, does not mean the whole world can. This is blocked by the firewall.
(unless you by mistake manually added a rule to iptables to allow it)
 
@kamoj
I don't use any of the supervision functions of the add-on. Do I need this cron job
Code:
/tmp/addons/cron/cron_supervision/crontabs:
* * * * * sleep 11 && /usr/bin/restart_supervision.sh
I see a little increase of CPU usage of the R7800 (I monitor CPU usage with top2 very closely after the issue I had a few weeks ago). Every 5-th and 7-th second I see the load going to 20-30% (spikes for a second). I suppose this is due to the new supervision functions (I may be wrong here) of the add-on. I didn't have those CPU load spikes around month ago. Yesterday before upgrading to the latest @Voxel firmware version I've uninstalled the add-on and I've monitored the CPU load with 1.0.2.83SF and 1.0.2.84SF but the CPU load was merely 1-3%.
 
@kamoj
I don't use any of the supervision functions of the add-on. Do I need this cron job
Code:
/tmp/addons/cron/cron_supervision/crontabs:
* * * * * sleep 11 && /usr/bin/restart_supervision.sh
I see a little increase of CPU usage of the R7800 (I monitor CPU usage with top2 very closely after the issue I had a few weeks ago). Every 5-th and 7-th second I see the load going to 20-30% (spikes for a second). I suppose this is due to the new supervision functions (I may be wrong here) of the add-on. I didn't have those CPU load spikes around month ago. Yesterday before upgrading to the latest @Voxel firmware version I've uninstalled the add-on and I've monitored the CPU load with 1.0.2.83SF and 1.0.2.84SF but the CPU load was merely 1-3%.
I might be wrong, but I think you are barking up the wrong tree.

If you don't use any of the restart supervision function, it won't cause any cpu load.

If you have not enabled any restart supervision function the add-on will just ping twice, once per minute. Not every 5th / 7th second!
Once to see if internet is accessible, and once to see if DNS is working.
That's all. Ping is not cpu heavy, so it won't use any cpu to talk about.

If you want to disable it you can change the "once per minute" to your liking.
E.g. to run the supervision only once a year at new year:
Code:
nvram set kamoj_supervision_update_time="0 0 1 1 *"; nvram commit

Also to measure cpu load in a system that is not loaded to 100% is very difficult.
E.g. the top/ps/netdata etc consumes a lot of cpu themselves.
This "art" is worth a topic itself, and much is written on the topic!
I.e. you can't measure without effecting the cpu load.
The more you measure, the higher load!

It would be interesting to know how you do! Can you describe your way please?

The add-on also has several other "daemons" running that can cause load, e.g. measure the temperatures in the router.
These daemons can be listed using the add-on function: System Information: Addon processes: show
The name of the .pid-file gives a clue of what the daemon is doing.

Since all add-on functions are implemented as non-compiled code, you can review all code, and change to your likings.

Maybe you can study the interrupt distribution as well, or I/O-processes.
There is much fun going on "under the hood".

Thank you , and and please report back your findings!
 
Thank you very good to get your log files.

Please enable functions/settings one by one until you find the culprit - if there really is any problem.
This is the least I expect from a beta tester.

You have some settings enabled that disturb your measurements:
  • Aegis. Disable Aegis while doing these measurements.
  • Set time stamps on dmesg logs. This setting also makes Aegis to bug out and stop logging after some 9000 seconds from boot!
    This Netgear bug is discussed elsewhere, so just switch this option off if you use Aegis.
  • Router Information. Switch off auto update (every 20 seconds), and don't make manual refresh.
    Router Information consumes "much" cpu when run, as it collect and calculates "thousands" of measurements.
  • Save logs. This function also consumes much cpu and disk i/o.
    It also writes a lot to disk and may effect memory. Switch it off, until you know what you do/want it for.
  • If you don't use DLNA, switch it off!
  • If you don't use WPS, switch it off!
  • You seems to have some error with your routers memory.
    (msm_nand_read_oob 0 800 0 + msm_nand_read_oob c80000 800 0 failed)
    Please provide the "flash info" output!
Apart from all your settings causing artificial cpu load I must ask:
Do you experience any problem?
Any function that is not working?
@kamoj Thankfully for your guides, since version V5.4b20 I don’t have any issues with the addon or the router!

Apologises for the very late response, but I have been busy with my little girl and I don’t have the time to sit down and to reply.

All my issues been resolved by disabling the “Save the system logs at boot time”, after I disabled I had no more high CPU usage and AdguradHome also was working without any crashes. In meantime I disabled all the unnecessary options like WPS,ReadyCloud,Kwilt etc. Couple of days ago I updated to Voxel latest firmware V1.0.2.84SF and to your latest addon V5.4b26, everything working as expected.

Big thanks to you @kamoj for your amazing addon and also for @Voxel because he is providing for us a superb firmware, what is letting us to do all this stuff with the router, last but not least, thank you to @HELLO_wORLD for great firewall blocklist script! Netgear should offer a job you guys!

Keep up your excellent work and stay safe!
 
@kamoj Thankfully for your guides, since version V5.4b20 I don’t have any issues with the addon or the router!

Apologises for the very late response, but I have been busy with my little girl and I don’t have the time to sit down and to reply.

All my issues been resolved by disabling the “Save the system logs at boot time”, after I disabled I had no more high CPU usage and AdguradHome also was working without any crashes. In meantime I disabled all the unnecessary options like WPS,ReadyCloud,Kwilt etc. Couple of days ago I updated to Voxel latest firmware V1.0.2.84SF and to your latest addon V5.4b26, everything working as expected.

Big thanks to you @kamoj for your amazing addon and also for @Voxel because he is providing for us a superb firmware, what is letting us to do all this stuff with the router, last but not least, thank you to @HELLO_wORLD for great firewall blocklist script! Netgear should offer a job you guys!

Keep up your excellent work and stay safe!
:) Thank you for you report, and I must say I share your appreciation of the fantastic job @Voxel and @HELLO_wORLD are doing!
 
Hi,

I have a r7800 and would like to get the BETA for Kamoj.
Used the older version a while ago, switched back to voxel native but now I have a another router and want to use the r7800 as 2nd one.
Been using Voxel's since the early days btw..... happy user.

I work in IT, web / dbms environments.

Regards,

Marachi.
 
As I wrote, these are small issues. Thanks again for all your efforts. M.
@stern67,

As you described, on my R7800 device in router mode (both OpenVPN and Wireguard clients disabled), the Internet (WAN) LED also blinks amber some time after a reboot and your trick of unplugging and plugging back in the WAN cable also works for me in changing it to flashing white.

It appears to be a flaw with Netgear's logic related to Traffic control (ref. user's manual, page 92):

Code:
In the Traffic Control section, specify whether the router should issue a warning message before the monthly
limit of Mbytes or hours is reached.
By default, the value is 0 and no warning message is issued.You can select one of the following to occur when
the limit is attained:
• The Internet LED blinks white or amber.
• The Internet connection is disconnected and disabled.

I have my traffic meter and traffic control settings all disabled.

Traffic_Control.JPG


Also, as you have indicated, the WAN connection speed is still listed 1000M/Full, even when it's flashing amber, and everything is working normally.
 
@stern67,

As you described, on my R7800 device in router mode (both OpenVPN and Wireguard clients disabled), the Internet (WAN) LED also blinks amber some time after a reboot and your trick of unplugging and plugging back in the WAN cable also works for me in changing it to flashing white.

Also, as you have indicated, the WAN connection speed is still listed 1000M/Full, even when it's flashing amber, and everything is working normally.
Thank you for the reporting, both of you!
I've found an error in the add-on when you don't use VPN. So I hope it's working in next release!
To get light back to normal you can run:

Code:
led_all normal
or
Code:
ledcontrol -n wan -c green -s on
You can add the command(s) in the quite new add-on function Settings 2: Create start script entries manually (Executed at boot time) + press "Save start script"
 
Hello everyone,
I am having issues installing Entware to my USB stick with entware_install.sh command by Kamoj. I can see the directories created, but Entware wont work. (ie. cannot install netdata)
I format the USB to FAT32 with my PC, then connect it to USB1 of my R7800, reboot router and run entware_install.sh via SSH and I get the errors of the log file i have attached. (I have removed some tar lines before the end, otherwise it wouldn't let me upload the file)
I read on another thread there was an issue to this command, but have been fixed since then. Maybe the command is broken again?
Can you please take a look? Many thanks in advance!
I also get the same issue using the latest beta 5.4 b26. I heard it was working, so I took my currently formatted NTFS drive in hopes it'd format the drive to EXT4 for me. Still downloaded the files and created/mounted swapfile, but it's like the script entware_install.sh doesn't know how to forcibly dismount a drive, format it proper and given the right label. This also doesn't seem to take into account different starting labels and mountpoints like /mnt/sdf1 or mnt/sdb1, crashes much the same way as this user's log shows.
 
I also get the same issue using the latest beta 5.4 b26. I heard it was working, so I took my currently formatted NTFS drive in hopes it'd format the drive to EXT4 for me. Still downloaded the files and created/mounted swapfile, but it's like the script entware_install.sh doesn't know how to forcibly dismount a drive, format it proper and given the right label. This also doesn't seem to take into account different starting labels and mountpoints like /mnt/sdf1 or mnt/sdb1, crashes much the same way as this user's log shows.
Thank you for your report.
I suppose you have followed @Voxel instructions and I hope that you can confirm here that it is working?!

Within in this beta program, you are supposed to help find not only errors, but also help, pinpoint, investigate, give solutions etc.

Since I don't have any record of you being invited to the beta testing, may I ask you how you got hold of the v5.4 b26?

PS:
Thank you Kamoj! Formating USB partition to ext4 and naming it "optware" solved the issue.
 
Last edited:

Latest threads

Sign Up For SNBForums Daily Digest

Get an update of what's new every day delivered to your mailbox. Sign up here!
Top