pixelserv pixelserv - A Better One-pixel Webserver for Adblock

[jrmwvu04]

The tablets we just have at home for entertaining the two-year-old and misc browsing are just swipe-to-unlock; so they may be the source of most of these errors.

Mobile apps for young children are among the most aggressive I’ve ever had on my network, especially in the realm of tracking and metrics.

 

[kvic]

Create a discord channel and keep it invite only.

Thanks. I created a channel for testing libc. I'm totally new to Discord but for the brief moment I tried, sounds like lot of fun.

Same user id, kvic!

 

[Asad Ali]

Thanks. I created a channel for testing libc. I'm totally new to Discord but for the brief moment I tried, sounds like lot of fun.

Same user id, kvic!

Great, now you'll need to manually send your channel invite links to whom you want to be a part of it, we can't join on our own.

 

[kvic]

Sure, sounds good! Would like to figure this out.... although I may already have cracked the nut. You can't have a user certificate if you don't have a screen lock password/pin/pattern, apparently. The tablets we just have at home for entertaining the two-year-old and misc browsing are just swipe-to-unlock; so they may be the source of most of these errors. What do you think?

Can no longer add ppl to the original group due to forum change. But I sent you the essential info.

 

[kvic]

Great, now you'll need to manually send your channel invite links to whom you want to be a part of it, we can't join on our own.

lol..bear with a newbie. I created a "server" named "pixelserv-tls". Inside the server, currently have two channels, one default and one I created for "testing libc entware"

Not sure how popular Discord among folks here. I guess ppl could try to ping me over there..

 

[Asad Ali]

lol..bear with a newbie. I created a "server" named "pixelserv-tls". Inside the server, currently have two channels, one default and one I created for "testing libc entware"

Not sure how popular Discord among folks here. I guess ppl could try to ping me over there..

We can't find/join your server without an invite link.

Here's a little guide for you:

https://support.discordapp.com/hc/en-us/articles/204155938-How-do-I-invite-friends-to-my-server-

 

[kvic]

We can't find/join your server without an invite link.

Here's a little guide for you:

https://support.discordapp.com/hc/en-us/articles/204155938-How-do-I-invite-friends-to-my-server-

alright...let's try: https://discord.gg/7arCS

The link is valid for 1 hr.

 

[Asad Ali]

alright...let's try: https://discord.gg/7arCS

The link is valid for 1 hr.

I'm in!!

 

[kvic]

I've sent out invitations to join the Discord channel for "testing - armv7 libc Entware".

Invitations are based on your interest expressed on this issue in this thread. If I've missed anyone, pls let me know.

The fun part shall start in the next 12 hrs. But essential thing to test is perhaps in a day or two.

 

[kvic]

A fresh cycle of another near 38 hours..on 2.1.3-test.4

 

[quant88]

With 2.1.3-test.4, I recorded

1) Quite a number of "Reached max retries"
2) Log stating "server(null)"

Jul 18 22:54:53 pixelserv-tls[19766]: ib.adnxs.com _.adnxs.com missing
Jul 18 22:54:54 pixelserv-tls[19766]: ads.pubmatic.com _.pubmatic.com missing
Jul 18 23:17:49 pixelserv-tls[19766]: handshake failed: client 192.168.1.24:53290 server (null). Lib(20) Func(118) Reason(252)
Jul 18 23:20:18 pixelserv-tls[19766]: handshake failed: client 192.168.1.24:53354 server (null). Lib(20) Func(118) Reason(252)
Jul 18 23:42:27 pixelserv-tls[19766]: handshake failed: reached max retries. client 192.168.1.203:57625 server lepodownload.mediatek.com
Jul 18 23:43:39 pixelserv-tls[19766]: cdn-akamai.mookie1.com _.mookie1.com missing
Jul 18 23:49:47 pixelserv-tls[19766]: handshake failed: reached max retries. client 192.168.1.82:40321 server proton.flurry.com
Jul 18 23:49:47 pixelserv-tls[19766]: handshake failed: client 192.168.1.82:57401 server (null). Lib(20) Func(138) Reason(267)
Jul 19 00:12:16 pixelse

Instance 1 : Mostly clean requests.

pixelserv-tls 2.1.3-test.4 (compiled: Jul 17 2018 01:44:19) options: 192.168.1.2 -l 2 -c 700
uts 2d 00:41 process uptime
log 2 critical (0) error (1) warning (2) notice (3) info (4) debug (5)
kcc 3 number of active service threads
kmx 56 maximum number of service threads
kvg 1.86 average number of requests per service thread
krq 1963 max number of requests by one service thread
req 13978 total # of requests (HTTP, HTTPS, success, failure etc)
avg 1915 bytes average size of requests
rmx 31823 bytes largest size of request(s)
tav 35 ms average processing time (per request)
tmx 3168 ms longest processing time (per request)
slh 5341 # of accepted HTTPS requests
slm 358 # of rejected HTTPS requests (missing certificate)
sle 0 # of rejected HTTPS requests (certificate available but bad)
slc 2740 # of dropped HTTPS requests (client disconnect without sending any request)
slu 248 # of dropped HTTPS requests (other TLS handshake errors)
uca 0 slu break-down: # of unknown CA reported by clients
uce 4 slu break-down: # of unknown cert reported by clients
ush 51 slu break-down: # of shutdown by clients after ServerHello

Instance 2 : All known suspicious requests are all redirected to this instance.

pixelserv-tls 2.1.3-test.4 (compiled: Jul 17 2018 01:44:19) options: 192.168.1.3 -c 200 -O 300 -l 1

uts 2d 00:53 process uptime
log 1 critical (0) error (1) warning (2) notice (3) info (4) debug (5)
kcc 21 number of active service threads
kmx 40 maximum number of service threads
kvg 1.63 average number of requests per service thread
krq 4633 max number of requests by one service thread
req 708643 total # of requests (HTTP, HTTPS, success, failure etc)
avg 1873 bytes average size of requests
rmx 133746 bytes largest size of request(s)
tav 26 ms average processing time (per request)
tmx 10006 ms longest processing time (per request)
slh 9927 # of accepted HTTPS requests
slm 259 # of rejected HTTPS requests (missing certificate)
sle 0 # of rejected HTTPS requests (certificate available but bad)
slc 9542 # of dropped HTTPS requests (client disconnect without sending any request)
slu 660660 # of dropped HTTPS requests (other TLS handshake errors)
uca 55 slu break-down: # of unknown CA reported by clients
uce 655745 slu break-down: # of unknown cert reported by clients
ush 221 slu break-down: # of shutdown by clients after ServerHello

 

[kvic]

@quant88

What a "Instance 2"

Instance 1 has two failures (252 & 267) that do not have server name. It's expected behaviour. Server name is only available when the handshake can somewhat go through in the middle of the process. If they fail early, we can't capture server names. That's what happened.

Roughly what percentage of "reached max retries" on a daily basis? I think my "tls report script" can help to get that stats.

 

[kvic]

my "tls report script" can help to get that stats.

today's stats. "reached max retries" is less than 2% out of total failures in the past 24 hours. Perhaps we shall have # of retries to be a command line option..let me think through it.

 

[quant88]

@kvic

Nice script
Let me go through to set this up.

today's stats. "reached max retries" is less than 2% out of total failures in the past 24 hours. Perhaps we shall have # of retries to be a command line option..let me think through it.

 

[kvic]

Some update on this

I've been trying to find some benchmarks to test the new feature. Luckily got one after a few days. Ported the bench test to 56U and run. Bad news: performance is worse than without the feature!

Run the same bench test on Debian Jessie (very outdated software components) vs Arch Linux (everything is latest and..greatest). Same hardware. The new feature (together with many other new features I assume) shows significant boost in the bench test!

A few obvious questions:
1) my porting of the feature to 56U has problem
2) the bench test isn't appropriate to this particular new feature
3) etc..

It requires more of my time to look at them. Perhaps another week (if not more)..

p.s. there is an early reply to this thread that's pending moderation. People can read here for the time being.

This test is going very well. People in general see a ~44% performance boost on memory allocation benchmark. That is going to benefit pixelserv-tls (heavily multi-threads with many malloc operations) but in general it benefits all software downloaded from Entware.

Now is open for more testers! Interested ppl..pls raise your hands. I'll send out Discord invites!

 

[penguin22]

You can include me; as you know, I have a high use network with 30+ devices.

This test is going very well. People in general see a ~44% performance boost on memory allocation benchmark. That is going to benefit pixelserv-tls (heavily multi-threads with many malloc operations) but in general it benefits all software downloaded from Entware.

Now is open for more testers! Interested ppl..pls raise your hands. I'll send out Discord invites!

 

[OMNI619]

Can you emailed to me please I would like to test it I have about 35 devices my network gets hammered thanks

Sent from my SAMSUNG-SM-G920AZ using Tapatalk

 

[Goobs]

Me too! With 44 devices on the network I would love to test it since you mentioned it is helping with memory usage!!!

 

[tomsk]

Hi... having a small issue with the SNB website after a fresh install of of pixelserv via the Absolution script.
Once i import the pixelserv ca.crt the blank spaces where the ads would normally go now have copies of the top of the page in addition the the page grinding to a halt..... if i remove the cert and restart the browser the ad spaces become blank again. Pixelserv version is 2.1.1
Im using a Mac and the problem shows in both safari and chrome.

 

[kvic]

Hi... having a small issue with the SNB website after a fresh install of of pixelserv via the Absolution script.
Once i import the pixelserv ca.crt the blank spaces where the ads would normally go now have copies of the top of the page in addition the the page grinding to a halt..... if i remove the cert and restart the browser the ad spaces become blank again. Pixelserv version is 2.1.1
Im using a Mac and the problem shows in both safari and chrome.

ABS users figured out some domains to whitelist to resolve the issue. Shouldn't by default it's already whitelisted in ABS with a fresh install? Hmm..

General recommendation for troubleshooting:

Open browser's Web Inspector and turn on pixelserv-tls logging (instructions here). Load the problematic page in browser.

Perhaps you will see what domains give you trouble..