Diversion Diversion - the Router Ad-Blocker

[Khadanja]

Or you coud use the top command in the shell to see what process is using the cpu.
Or even better install htop:
opkg install htop

Here is a screen shot from htop

 

[Khadanja]

That's a very good question. I'm not sure about the USB drive so it's best to wait for someone with a little more experience to answer if you prefer keeping it with all its current data.

If you don't care and want to truly start over from scratch (which is what I did), nuke everything and proceed (i.e. formatting USB drive, resetting router, manually re-installing all the services you want).[/QUOTE

Well this is what I did, nuked everything and started from scratch. High RAM and CPU starts after installing services. Can you post your top or htop result?

 

[thelonelycoder]

Here is a screen shot from htop

That's the webserver for the UI, running on https that's using so much cpu power to show you the WebUI pages.
Nothing abnormal IMHO.

 

[Khadanja]

That's the webserver for the UI, running on https that's using so much cpu power to show you the WebUI pages.
Nothing abnormal IMHO.

So why is it not using high cpu for others like @#TY

 

[#TY]

Well this is what I did, nuked everything and started from scratch. High RAM and CPU starts after installing services. Can you post your top or htop result?

Sure, here you are. It fluctuates obviously but its very low CPU usage most of the time.

 

[thelonelycoder]

So why is it not using high cpu for others like @#TY

Why bother, the spikes probably come from when pages are reloaded. Once logged out they disappear.

 

[Khadanja]

ok nuked everything and started again, looks lot better. Can anyone tell me if installation of amtm,skynet,diversion can be backed up or what will happen if router is reset or NVRAM reset using WPS button? Will I have to install again manually if router reset or NVRAM reset?

 

[bigmog]

Running Diversion 4.0.7 on Merlin 384.9 on Asus AC86U

After about a week or so of continued operation, my internet connection is failing. When this happens, I check the system log and the timing lines up with events related to diversion weekly updates

Mar 30 02:00:00 rc_service: service 25078:notify_rc restart_dnsmasq
Mar 30 02:00:00 custom_script: Running /jffs/scripts/dnsmasq.postconf (args: /etc/dnsmasq.conf ) - max timeout = 120s
Mar 30 02:00:00 Diversion: restarted Dnsmasq to apply settings, from /jffs/scripts/dnsmasq.postconf
Mar 30 02:00:11 Diversion: sorting of /opt/share/diversion/list/blockinglist file failed, using previous file, from /opt/share/diversion/file/update-bf.div
Mar 30 02:00:12 Diversion: updated Medium blocking list from 2 hosts files, 264341 domains are now blocked, from /opt/share/diversion/file/update-bf.div
Mar 30 02:00:13 Diversion: updated and reset weekly ads counter: 544,009 total, 40,366 this week, 3,584 new since last count, from /opt/bin/diversion
Mar 30 02:00:13 Diversion: reset dnsmasq log files (weekly cron job), from /opt/share/diversion/file/update-bf.div
Mar 30 02:00:46 rc_service: watchdog 792:notify_rc start_dnsmasq
Mar 30 02:00:46 custom_script: Running /jffs/scripts/dnsmasq.postconf (args: /etc/dnsmasq.conf ) - max timeout = 120s
Mar 30 02:00:47 Diversion: restarted Dnsmasq to apply settings, from /jffs/scripts/dnsmasq.postconf

To troubleshoot this further, I attempt to run diversion but get this error.

dg@zelda:/tmp/home/root# diversion
/opt/bin/diversion: line 233: can't fork​

Thinking this is a memory exhaustion issue, I check the available memory with top but find no such issue.

Mem: 300024K used, 140396K free, 1520K shrd, 760K buff, 96612K cached
CPU: 0.0% usr 0.0% sys 0.0% nic 100% idle 0.0% io 0.0% irq 0.0% sirq
Load average: 2.98 3.03 3.03 1/127 5842

This has happened perhaps 4 times so far and the only resolution I have is to reboot the router. Then all is well for another week.

pixelserv-tls is enabled.
swap file is not.
blocklist level is medium

Any ideas?

 

[L&LD]

ok nuked everything and started again, looks lot better. Can anyone tell me if installation of amtm,skynet,diversion can be backed up or what will happen if router is reset or NVRAM reset using WPS button? Will I have to install again manually if router reset or NVRAM reset?

It seems to me like you're creating the issues you're seeing over and over? Are you still using a backup config file as you posted a few posts ago? Are you just doing a partial M&M Config? As to the use of backing up the scripts, I would say unless you know how to solve any possible issues there, don't. There have been many updates to the scripts in the last few weeks and a fresh install of all of them is what solved many issues for many people here (too many posts to look them all up now).

The high CPU usage you're seeing is normal. On some routers, it lasts a few minutes, on some I've seen it last on and off for over an hour. This is after a full M&M Config and before applying any features or scripts. I attribute this to the network/client specifics, the quality of the flash used in different models and the other hardware/software capabilities in each (CPU, RAM, drivers, package versions, etc.). This is not based on examination; this is just my informed, best guess.

Use the links in my signature to bring your router to a good/known state.

After doing so, install amtm onto a freshly formatted jffs partition.

Plug in the USB drive and format it to Ext4 with journalling.

After it has rebooted, create a swap file (I use 2GB for all router installs) using amtm. Also, enable the amtm disk check script.

Install Diversion and let it install Entware and pixelserv-tls.

Check for updates to Entware in the Diversion menu; ep, 6, 6. After it finishes, exit to amtm.

Install the rest of the scripts you want/need. Testing each script and the preceding scripts previously installed before moving onto the next install. Do the same for any needed features in the router too now. Testing all previously installed scripts and enabled features as you go.

When all are installed, I would reboot the router (via the GUI) and check that the disk checker log 'passed'.

Leave the router to settle down (you can just use the rest of your network normally) for at least an hour. I would perform another GUI reboot and check the disk checker log once more.

After a suitable time from the last reboot has passed, the router's CPU's will be mostly idling if your network is idling too.

As for the RAM usage? Let it be used. That is why we installed the swap file for the router to use.

The more RAM that is utilized, the more responsive the router and the network will be.

As a matter of fact, if your router has it, disable the "Memory Management: Regularly flush caches (default: Yes)" and set it to 'No'. You can find this in the Tools, Other Settings, Advanced Tweaks and Hacks section in the GUI.

 

[dave14305]

Running Diversion 4.0.7 on Merlin 384.9 on Asus AC86U

After about a week or so of continued operation, my internet connection is failing. When this happens, I check the system log and the timing lines up with events related to diversion weekly updates

Mar 30 02:00:00 rc_service: service 25078:notify_rc restart_dnsmasq
Mar 30 02:00:00 custom_script: Running /jffs/scripts/dnsmasq.postconf (args: /etc/dnsmasq.conf ) - max timeout = 120s
Mar 30 02:00:00 Diversion: restarted Dnsmasq to apply settings, from /jffs/scripts/dnsmasq.postconf
Mar 30 02:00:11 Diversion: sorting of /opt/share/diversion/list/blockinglist file failed, using previous file, from /opt/share/diversion/file/update-bf.div
Mar 30 02:00:12 Diversion: updated Medium blocking list from 2 hosts files, 264341 domains are now blocked, from /opt/share/diversion/file/update-bf.div
Mar 30 02:00:13 Diversion: updated and reset weekly ads counter: 544,009 total, 40,366 this week, 3,584 new since last count, from /opt/bin/diversion
Mar 30 02:00:13 Diversion: reset dnsmasq log files (weekly cron job), from /opt/share/diversion/file/update-bf.div
Mar 30 02:00:46 rc_service: watchdog 792:notify_rc start_dnsmasq
Mar 30 02:00:46 custom_script: Running /jffs/scripts/dnsmasq.postconf (args: /etc/dnsmasq.conf ) - max timeout = 120s
Mar 30 02:00:47 Diversion: restarted Dnsmasq to apply settings, from /jffs/scripts/dnsmasq.postconf

To troubleshoot this further, I attempt to run diversion but get this error.

dg@zelda:/tmp/home/root# diversion
/opt/bin/diversion: line 233: can't fork​

Thinking this is a memory exhaustion issue, I check the available memory with top but find no such issue.

Mem: 300024K used, 140396K free, 1520K shrd, 760K buff, 96612K cached
CPU: 0.0% usr 0.0% sys 0.0% nic 100% idle 0.0% io 0.0% irq 0.0% sirq
Load average: 2.98 3.03 3.03 1/127 5842

This has happened perhaps 4 times so far and the only resolution I have is to reboot the router. Then all is well for another week.

pixelserv-tls is enabled.
swap file is not.
blocklist level is medium

Any ideas?

Could be too many processes running, which would be unusual. What do you see if you run ps on the router?

 

[Zork]

What’s everyone CPU usage like on 68U after installing diversion? Mine spiked up after installing amtm then diversion & skynet. Average is around 68%. Is that normal? Spikes to 89-90% sometimes.

Have the same issue with amtm, only diversion is ok. But plus skynet cpu always on 80-90%

 

[L&LD]

Have the same issue with amtm, only diversion is ok. But plus skynet cpu always on 80-90%

Please see my reply to Khadanja above. Some of the suggestions may also help here too.

 

[bigmog]

Could be too many processes running, which would be unusual. What do you see if you run ps on the router?

After a recent reboot, 112 running processes.

dg@zelda:/tmp/home/root# ps | wc -l
112


I'll track this and see how this number grows over this week.

 

[skeal]

Will I have to install again manually if router reset or NVRAM reset?

In most cases yes. On some models you can choose to not wipe the jffs. On my AX88U there is no such option, when you reset to defaults in the webui the jffs gets wiped, again on this model. On my AC68U and AC3100 the option to not wipe the jffs is there. You can choose to backup the jffs on the save settings page, separate from the router config. This means you can fully restore your jffs, so a default reset is ok to do.

 

[#TY]

At the risk of asking too much lol, do you guys think it might be possible for Diversion to readjust the page after removing the ads?

 

[skeal]

At the risk of asking too much lol, do you guys think it might be possible for Diversion to readjust the page after removing the ads?

That would be controlled at the browser level, not by diversion. Sorry mate!

 

[#TY]

That would be controlled at the browser level, not by diversion. Sorry mate!

LOL damn....was worth a shot

 

[thelonelycoder]

Mar 30 02:00:11 Diversion: sorting of /opt/share/diversion/list/blockinglist file failed, using previous file, from /opt/share/diversion/file/update-bf.div

During the weekly blocking list update, your router runs out of memory. Enable the swap file, make it as large as the routers ram. For the 86U its 512MB.

 

[L&LD]

During the weekly blocking list update, your router runs out of memory. Enable the swap file, make it as large as the routers ram. For the 86U its 512MB.

I would appreciate if you could give me the reasons behind having the swap file as large as the RAM is. I have seen it repeated multiple times and I really don't know why?

I always create the largest available swap file with amtm on an Ext4 partition with journalling enabled (currently 2GB limit). I do this for any router I install it on. I see no harm or performance degradation as the USB drives used are at least 4GB and the scripts are very tiny, relatively!

Is there a real reason to only match the RAM for the swap file size?

 

[thelonelycoder]

Is there a real reason to only match the RAM for the swap file size?

Rule of thumb, really. 256MB would work for most. But I always set it at the routers RAM.
The RT-AX88U has 1GB RAM which is nice but somehow it still gets used, so I make the swap file equal to it.