pixelserv pixelserv - A Better One-pixel Webserver for Adblock

[kvic]

@rromeroa

Good to have things clarified and it looks like you know what you're doing

From your log, I could guess pixelserv-tls crashed very early on. None of the service threads expire & exit themselves. We could narrow down to the burst of six POST requests. I need to find an android client or some other way to simulate the situation. Tapatalk seems making requests differently on iOS.

 

[kvic]

@rromeroa

I spent more time trying to reproduce a crash with Tapatalk on Android. Still no luck. Upon launch of Tapatalk, my log is a bit different from yours. I have 3 POST each to a different ad hosts:

Dec  1 14:44:46 pixelserv-tls[769]: no file extension /settings from path /spi/v2/platforms/android/apps/com.quoord.tapatalkpro.activity/settings
Dec  1 14:44:46 pixelserv-tls[769]: 192.168.1.113 settings.crashlytics.com GET /spi/v2/platforms/android/apps/com.quoord.tapatalkpro.activity/settings?icon_hash=896eddadcef8635b3268053fb94171ce5f6627c4&display_version=7.1.8&source=4&instance=d4cd971161250eb2e29b76d1a077eb4bf826615e&build_version=971 HTTP/1.1 secure
Dec  1 14:44:46 pixelserv-tls[769]: POST socket: 13 Content-Length: 8528
Dec  1 14:44:46 pixelserv-tls[769]: POST socket: 13 expect length: 8528
Dec  1 14:44:46 pixelserv-tls[769]: POST socket:13 recv length:8192; errno:0
Dec  1 14:44:46 pixelserv-tls[769]: POST socket:13 recv length:336; errno:0
Dec  1 14:44:46 pixelserv-tls[769]: 192.168.1.113 data.flurry.com POST /aap.do HTTP/1.1 secure
Dec  1 14:44:46 pixelserv-tls[769]: [-binary POST content not dumped-]
Dec  1 14:44:47 pixelserv-tls[769]: Exit recv loop socket:16 rv:0 errno:0 wait_cnt:63 num_req:1
Dec  1 14:44:56 pixelserv-tls[769]: POST socket: 14 Content-Length: 1044
Dec  1 14:44:56 pixelserv-tls[769]: POST socket: 14 expect length: 0
Dec  1 14:44:56 pixelserv-tls[769]: 192.168.1.113 ssl.google-analytics.com POST /batch HTTP/1.1 secure
Dec  1 14:44:56 pixelserv-tls[769]: [-binary POST content not dumped-]
Dec  1 14:44:56 pixelserv-tls[769]: Exit recv loop socket:18 rv:0 errno:0 wait_cnt:93 num_req:1

Also tried to simulate 6 POST with binary data in a row. Still can't crash v2.0.0-rc1

From your log, you tried to access the servstats page? Did you see the page? How long after you saw the page pixelserv-tls crash? Also does it crash as fast as with and without '-l 5'? More details you could smell may help to investigate the issue.

 

[shooter40sw]

Hi ! I would like to install the updated pixelserv, but I installed it form ab-solutions, I read that we have to wait about 2 weeks or more, is there any other way to do this so I can install pixelserv without disrupting any settings within ab-solution or pixelserv? thanks

 

[Butterfly Bones]

I have something odd. When I use the current Kk release I see certs added via the syslog. When I install the 2.0-rc, I no longer see certs added, even if I add the "-l 5" switch. With Kk starting and stopping AB-Solution is quick; with 2.0-rc or any of the tests, AB-Solution takes twice as long to load and 20-30 seconds to exit. With the tests or -rc versions, I see pixelserv restart when I enter and exit AB-S. That does not occur with AB-S using Kk unless I specifically use the ps then 4 (restart) menu options in AB-S.

The other odd thing is permissions, some with "nobody root" and some with "1000 1000" and most certs start with _.xxx (underscore dot cert but some are just .xxx (dot cert). I just cleared the certs and generated a new ca.crt back on Nov. 24th. Here is a list. Not sure if this is changes on different versions of Kk and the tests / -rc? I do remember thelonelycoder making this change to "nobody" in AB-S a few weeks back. I tried to find that reference but SNB does not like "nobody".

Search Results for Query: nobody

1. The search could not be completed because the search keywords were too short, too long, or too common.

drwxrwxr-x    2 nobody   1000          4096 Nov 30 18:58 .
drwxrwxr-x    3 1000     1000          4096 Nov 28 19:45 ..
-rw-rw-r--    1 1000     1000          1580 Nov 28 19:45 _.2mdn.net
-rw-rw-r--    1 1000     1000          1600 Nov 28 19:45 _.4seeresults.com
-rw-rw-rw-    1 nobody   root          1588 Nov 30 11:47 _.addthis.com
-rw-rw-r--    1 1000     1000          1592 Nov 28 19:45 _.adobedtm.com
-rw-rw-rw-    1 nobody   root          1592 Nov 30 18:58 _.adobetag.com
-rw-rw-r--    1 1000     1000          1592 Nov 28 19:45 _.akamaihd.net
-rw-rw-r--    1 1000     1000          1612 Nov 28 19:45 _.amazon-adsystem.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.amazon.com
-rw-rw-r--    1 1000     1000          1600 Nov 28 19:45 _.appdynamics.com
-rw-rw-r--    1 1000     1000          1596 Nov 28 19:45 _.appsflyer.com
-rw-rw-r--    1 1000     1000          1596 Nov 28 19:45 _.at.atwola.com
-rw-rw-rw-    1 nobody   root          1584 Nov 30 13:13 _.bkrtx.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.bluekai.com
-rw-rw-r--    1 1000     1000          1592 Nov 28 19:45 _.boldchat.com
-rw-rw-rw-    1 nobody   root          1600 Nov 30 04:16 _.casalemedia.com
-rw-rw-rw-    1 nobody   root          1596 Nov 30 13:13 _.chartbeat.com
-rw-rw-r--    1 1000     1000          1604 Nov 28 19:45 _.civicscience.com
-rw-rw-r--    1 1000     1000          1592 Nov 28 19:45 _.cloudfront.net
-rw-rw-r--    1 1000     1000          1596 Nov 28 19:45 _.contextweb.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.criteo.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.criteo.net
-rw-rw-r--    1 1000     1000          1600 Nov 28 19:45 _.crittercism.com
-rw-rw-r--    1 1000     1000          1596 Nov 28 19:45 _.crwdcntrl.net
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.demdex.net
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.disqus.com
-rw-rw-r--    1 1000     1000          1596 Nov 28 19:45 _.doubleclick.net
-rw-rw-r--    1 1000     1000          1596 Nov 28 19:45 _.ensighten.com
-rw-rw-r--    1 1000     1000          1592 Nov 28 19:45 _.exelator.com
-rw-rw-rw-    1 nobody   root          1616 Nov 29 20:56 _.fls.doubleclick.net
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.flurry.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.forbes.com
-rw-rw-rw-    1 nobody   root          1608 Nov 30 05:51 _.foreseeresults.com
-rw-rw-r--    1 1000     1000          1604 Nov 28 19:45 _.g.doubleclick.net
-rw-rw-r--    1 1000     1000          1592 Nov 28 19:45 _.geotrust.com
-rw-rw-r--    1 1000     1000          1596 Nov 28 19:45 _.getclicky.com
-rw-rw-r--    1 1000     1000          1612 Nov 28 19:45 _.google-analytics.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.google.com
-rw-rw-r--    1 1000     1000          1612 Nov 28 19:45 _.googleadservices.com
-rw-rw-r--    1 1000     1000          1616 Nov 28 19:45 _.googlesyndication.com
-rw-rw-r--    1 1000     1000          1612 Nov 28 19:45 _.googletagmanager.com
-rw-rw-r--    1 1000     1000          1616 Nov 28 19:45 _.googletagservices.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.gstatic.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.hlserve.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.hubspot.com
-rw-rw-r--    1 1000     1000          1604 Nov 28 19:45 _.imrworldwide.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.indexww.com
-rw-rw-r--    1 1000     1000          1620 Nov 28 19:45 _.ingest.crittercism.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.kochava.com
-rw-rw-rw-    1 nobody   root          1576 Nov 30 13:15 _.krxd.net
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.maxmind.com
-rw-rw-r--    1 1000     1000          1584 Nov 28 19:45 _.media.net
-rw-rw-r--    1 1000     1000          1592 Nov 28 19:45 _.mixpanel.com
-rw-rw-rw-    1 nobody   root          1592 Nov 30 04:16 _.moatads.com
-rw-rw-rw-    1 nobody   root          1584 Nov 30 13:13 _.mxpnl.com
-rw-rw-r--    1 1000     1000          1592 Nov 28 19:45 _.newrelic.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.nytimes.com
-rw-rw-rw-    1 nobody   root          1596 Nov 30 13:13 _.optimizely.com
-rw-rw-r--    1 1000     1000          1596 Nov 28 19:45 _.optnmnstr.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.outbrain.com
-rw-rw-r--    1 1000     1000          1596 Nov 28 19:45 _.pubnation.com
-rw-rw-rw-    1 nobody   root          1588 Nov 30 07:04 _.pusher.com
-rw-rw-r--    1 1000     1000          1596 Nov 28 19:45 _.quantserve.com
-rw-rw-r--    1 1000     1000          1600 Nov 28 19:45 _.query.yahoo.com
-rw-rw-r--    1 1000     1000          1584 Nov 28 19:45 _.res-x.com
-rw-rw-r--    1 1000     1000          1580 Nov 28 19:45 _.rlcdn.com
-rw-rw-r--    1 1000     1000          1616 Nov 28 19:45 _.scorecardresearch.com
-rw-rw-rw-    1 nobody   root          1612 Nov 30 17:09 _.searchdimension.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.segment.io
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.servebom.com
-rw-rw-r--    1 1000     1000          1612 Nov 28 19:45 _.services.disqus.com
-rw-rw-r--    1 1000     1000          1604 Nov 28 19:45 _.stats.paypal.com
-rw-rw-r--    1 1000     1000          1600 Nov 28 19:45 _.theguardian.com
-rw-rw-r--    1 1000     1000          1604 Nov 28 19:45 _.thoughtleadr.com
-rw-rw-r--    1 1000     1000          1592 Nov 28 19:45 _.tinypass.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.tiqcdn.com
-rw-rw-rw-    1 nobody   root          1588 Nov 30 14:04 _.truste.com
-rw-rw-rw-    1 nobody   root          1588 Nov 30 04:14 _.twitter.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.typekit.net
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.univide.com
-rw-rw-r--    1 1000     1000          1596 Nov 28 19:45 _.uservoice.com
-rw-rw-r--    1 1000     1000          1604 Nov 28 19:45 _.utilitooltech.com
-rw-rw-r--    1 1000     1000          1584 Nov 28 19:45 _.walmart.com
-rw-rw-r--    1 1000     1000          1608 Nov 28 19:45 _.washingtonpost.com
-rw-rw-r--    1 1000     1000          1600 Nov 28 19:45 _.wfxtriggers.com
-rw-rw-r--    1 1000     1000          1596 Nov 28 19:45 _.wpdigital.net
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 _.yieldmo.com
-rw-rw-r--    1 1000     1000           802 Nov 28 19:45 ca.crt
-rw-rw-r--    1 1000     1000           887 Nov 28 19:45 ca.key
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 disqusads.com
-rw-rw-r--    1 1000     1000          1580 Nov 28 19:45 etahub.com
-rw-rw-r--    1 1000     1000          1600 Nov 28 19:45 floodprincipal.com
-rw-rw-r--    1 1000     1000          1588 Nov 28 19:45 tororango.com

 

[tom-]

@tom-

Please keep it running as long as you can on 8a for this test. Touch wood if it get stuck again or you feel gd enough & bored, simply jump to rc1 & try with 10s select_timeout.

Well, roughly 16hrs later. Pixelserv-tls went un responsive again on 8a with the select_timeout of 10 seconds.
Here's the dumped stats:

daemon.crit pixelserv-tls[2094]: 58217 uts, 1 log, 15 kcc, 64 kmx, 1.06 kvg, 6 krq, 10865 req, 1546 avg, 62093 rmx, 73 tav, 3637 tmx, 25 slh, 15 slm, 0 sle, 7066 slc, 5 slu, 40 nfe, 0 gif, 11 ico, 61 txt, 14 jpg, 145 png, 0 swf, 50 sta, 0 stt, 191 ufe, 2 opt, 22 pst, 1 hed, 12 rdr, 0 nou, 0 pth, 0 204, 1 bad, 0 tmo, 7167 cls, 0 cly, 0 clt, 0 err

Guess, I'll test the select_timeout on the rc1 version.
rc1 becomes unresponsive in an even shorter amount of time than 8a with the select_timeout :

daemon.crit pixelserv-tls[11192]: 882 uts, 1 log, 1 kcc, 10 kmx, 1.70 kvg, 7 krq, 85 req, 253 avg, 608 rmx, 14 tav, 211 tmx, 0 slh, 0 slm, 0 sle, 26 slc, 0 slu, 0 nfe, 0 gif, 0 ico, 0 txt, 0 jpg, 9 png, 0 swf, 8 sta, 0 stt, 34 ufe, 0 opt, 0 pst, 0 hed, 0 rdr, 0 nou, 0 pth, 0 204, 0 bad, 0 tmo, 28 cls, 0 cly, 0 clt, 0 err

 

[Protik]

After running for about two days pixelserv became unresponsive.

Output of, killall -SIGUSR1 pixelserv-tls,

Dec  1 18:34:40 pixelserv-tls[1300]: 173120 uts, 1 log, 2 kcc, 40 kmx, 1.87 kvg, 29 krq, 9612 req, 1120 avg, 28644 rmx, 47 tav, 130882 tm

And output of grep pixelserv-tls /tmp/syslog.log,

Nov 26 09:04:06 pixelserv-tls[1301]: 84177 uts, 1 log, 10 kcc, 42 kmx, 6.74 kvg, 18443 krq, 24421 req, 1024 avg, 16489 rmx, 31 tav, 3312
Nov 26 09:04:06 admin: Started pixelserv-tls (AB-Solution) from .
Nov 26 09:04:06 pixelserv-tls[15461]: pixelserv-tls: v2.0.0-rc1 compiled: Nov 26 2017 20:53:17 options: 192.168.2.2
Nov 26 09:04:06 pixelserv-tls[15461]: Listening on :192.168.2.2:80
Nov 26 09:04:06 pixelserv-tls[15461]: Listening on :192.168.2.2:443
Nov 26 09:04:59 pixelserv-tls[15461]: 53 uts, 1 log, 3 kcc, 3 kmx, 1.50 kvg, 2 krq, 7 req, 685 avg, 1287 rmx, 73 tav, 352 tmx, 2 slh, 0 s
Nov 26 09:05:01 admin: Started pixelserv-tls (AB-Solution) from /tmp/mnt/pdas001/adblocking/addon/pixelserv-tls.add.
Nov 26 09:05:01 pixelserv-tls[16427]: pixelserv-tls: v2.0.0-rc1 compiled: Nov 26 2017 20:53:17 options: 192.168.2.2
Nov 26 09:05:01 pixelserv-tls[16427]: Listening on :192.168.2.2:80
Nov 26 09:05:01 pixelserv-tls[16427]: Listening on :192.168.2.2:443
Nov 27 10:35:45 pixelserv-tls[16427]: 91842 uts, 1 log, 1 kcc, 49 kmx, 1.71 kvg, 32 krq, 7461 req, 966 avg, 69690 rmx, 50 tav, 4511 tmx,
Nov 27 10:35:45 pixelserv-tls[8026]: pixelserv-tls: v2.0.0-rc1 compiled: Nov 26 2017 20:53:17 options: 192.168.2.2
Nov 27 10:35:45 pixelserv-tls[8026]: Listening on :192.168.2.2:80
Nov 27 10:35:45 pixelserv-tls[8026]: Listening on :192.168.2.2:443
Nov 27 10:35:46 admin: Started pixelserv-tls (AB-Solution) from .
Jul 31 17:00:38 pixelserv-tls[1300]: pixelserv-tls: v2.0.0-rc1 compiled: Nov 26 2017 20:53:17 options: 192.168.2.2
Jul 31 17:00:39 pixelserv-tls[1300]: Listening on :192.168.2.2:80
Jul 31 17:00:39 pixelserv-tls[1300]: Listening on :192.168.2.2:443
Jul 31 17:00:41 admin: Started pixelserv-tls (AB-Solution) from /jffs/scripts/services-start.
Dec  1 18:34:40 pixelserv-tls[1300]: 173120 uts, 1 log, 2 kcc, 40 kmx, 1.87 kvg, 29 krq, 9612 req, 1120 avg, 28644 rmx, 47 tav, 130882 tm

As usual, restarting the service resolve the issue.

From the logs I don't see anything going wrong. One reason for that might be that I did not change the log level or select_timeout.

To debug it step by step, I am changing log level to 5 for now to see if anything shows up in the log. After that I will increase select_timeout.

 

[kvic]

Hi ! I would like to install the updated pixelserv, but I installed it form ab-solutions, I read that we have to wait about 2 weeks or more

The correct way to estimate availability from Entware-ng: see #809.

is there any other way to do this so I can install pixelserv without disrupting any settings within ab-solution or pixelserv? thanks

v2.0.0-rc1, the last test build, is functionally equivalent to v2.0.0 release. You can use the beta install script to get rc1. Instructions on Github's frontpage. I paste below for your convenience:

Install beta version on Entware-ng systems

sh -c "$(wget -qO - https://kazoo.ga/pixelserv-tls/install-beta.sh)"

All these info are also mentioned in the release note on my blog post. More systematically organised info for each release available there. Please everyone help yourself & spend sometime there.

 

[kvic]

I have something odd. When I use the current Kk release I see certs added via the syslog. When I install the 2.0-rc, I no longer see certs added, even if I add the "-l 5" switch.

In version KK, "cert xxx generated & saved" is always output to syslog. Someone "complained" a few months ago that he didn't want to see such messages in syslog. Hence, in v2, it's changed as below:

• default log level is '-l 1' without CLI '-l xxx' option
• '-l 1' does not output "cert xxx generated & saved"
• '-l 2' (i.e. NOTICE) or above will output the message to syslog
• the message is output to syslog _only if_ a new certificate is generated by pixelserv-tls. Previous certificates are saved & re-used. Hence, when you run pixelserv-tls the longer the less likelihood you'll see this message in syslog even with '-l x' where x is >=2.

#FAQ#

With Kk starting and stopping AB-Solution is quick; with 2.0-rc or any of the tests, AB-Solution takes twice as long to load and 20-30 seconds to exit. With the tests or -rc versions, I see pixelserv restart when I enter and exit AB-S. That does not occur with AB-S using Kk unless I specifically use the ps then 4 (restart) menu options in AB-S.

@thelonelycoder you have some idea what might be not right..?

The other odd thing is permissions, some with "nobody root" and some with "1000 1000" and most certs start with _.xxx (underscore dot cert but some are just .xxx (dot cert). I just cleared the certs and generated a new ca.crt back on Nov. 24th. Here is a list. Not sure if this is changes on different versions of Kk and the tests / -rc? I do remember thelonelycoder making this change to "nobody" in AB-S a few weeks back. I tried to find that reference but SNB does not like "nobody".

To fix the permission manually now, you can type

chown -R nobody:root /opt/var/cache/pixelserv

I suspect your issue might be caused by a combination of AB-S and/or your action from shell.

Nevertheless, with v2, this issue will be non-issue as I've submitted a patch to Entware-ng.

 

[Martin_D]

Thank you for This release it is amazing, Keep it up.

pixelserv-tls: v2.0.0-rc1 compiled: Nov 26 2017 20:53:17 options: 10.82.235.7

uts    0d 01:15    process uptime
log    1    critical (0) error (1) warning (2) notice (3) info (4) debug (5)
kcc    3    number of active service threads
kmx    20    maximum number of service threads
kvg    4.53    average number of requests per service thread
krq    261    max number of requests by one service thread
req    1408    total # of requests (HTTP, HTTPS, success, failure etc)
avg    780 bytes    average size of requests
rmx    2355 bytes    largest size of request(s)
tav    4 ms    average processing time (per request)
tmx    207 ms    longest processing time (per request)
slh    0    # of accepted HTTPS requests
slm    22    # of rejected HTTPS requests (missing certificate)
sle    0    # of rejected HTTPS requests (certificate available but bad)
slc    276    # of dropped HTTPS requests (client disconnect without sending any request)
slu    38    # of dropped HTTPS requests (unknown error)
nfe    196    # of GET requests for server-side scripting
gif    2    # of GET requests for GIF
ico    130    # of GET requests for ICO
txt    27    # of GET requests for Javascripts
jpg    0    # of GET requests for JPG
png    0    # of GET requests for PNG
swf    0    # of GET requests for SWF
sta    146    # of GET requests for HTML stats
stt    0    # of GET requests for plain text stats
ufe    0    # of GET requests /w unknown file extension
opt    0    # of OPTIONS requests
pst    0    # of POST requests
hed    0    # of HEAD requests (HTTP 501 response)
rdr    7    # of GET requests resulted in REDIRECT response
nou    0    # of GET requests /w empty URL
pth    0    # of GET requests /w malformed URL
204    0    # of GET requests (HTTP 204 response)
bad    1    # of unknown HTTP requests (HTTP 501 response)
tmo    8    # of timeout requests (client connect w/o sending a request in 'select_timeout' secs)
cls    291    # of dropped requests (client disconnect without sending any request)
cly    0    # of dropped requests (client disconnect before response sent)
clt    0    # of dropped requests (reached maximum service threads)
err    0    # of dropped requests (unknown reason)

 

[thelonelycoder]

I have something odd. When I use the current Kk release I see certs added via the syslog. When I install the 2.0-rc, I no longer see certs added, even if I add the "-l 5" switch. With Kk starting and stopping AB-Solution is quick; with 2.0-rc or any of the tests, AB-Solution takes twice as long to load and 20-30 seconds to exit. With the tests or -rc versions, I see pixelserv restart when I enter and exit AB-S. That does not occur with AB-S using Kk unless I specifically use the ps then 4 (restart) menu options in AB-S.

I'll have to investigate that, when I have time.
Just make sure when you switch forth and back with pixelserv versions, only Kl and 2.0.0 understand the -l <loglevel> command.
Kk and earlier only accept -l, no log level given.

I can't find a reason why @kvic , but pixelserv-tls restarted tonight on its own, practically no workload and definitely no action by AB-Solution.
-l 3 is set, but the only entry is this:

Dec  1 22:23:50 pixelserv-tls[5057]: cert generated and saved: _.optnmnstr.com
Dec  2 05:10:42 tlc: AB-Solution created br0:pixelserv 192.168.2.2 via /jffs/scripts/wan-start
Dec  2 05:10:44 Skynet: [INFO] Startup Initiated... ( banmalware autoupdate usb=/tmp/mnt/entware )
Dec  2 05:10:47 ddns_update: ez-ipupdate: starting...

Interesting fact, Skynet ran its initialiation and my ddns updated shortly after.
The reason I noticed right away is because I have the routers LEDs turned off and this morning some of them were on. Which usually happens when WAN was down. But I see no WAN down / up in the syslog.

I do remember thelonelycoder making this change to "nobody" in AB-S a few weeks back. I tried to find that reference but SNB does not like "nobody".

AB sets the owner of the directory /opt/var/cache/pixelserv to "nobody", not the files contained in it.

 

[kvic]

I can't find a reason why @kvic , but pixelserv-tls restarted tonight on its own, practically no workload and definitely no action by AB-Solution.

pixelserv-tls cannot restart itself. So it should be caused by external force.

Or do you mean it crashed?

 

[thelonelycoder]

pixelserv-tls cannot restart itself. So it should be caused by external force.

Or do you mean it crashed?

I suspect a WAN event not registered in syslog then.
I'll try to recreate it later.

 

[pattiri]

I'm having below logs for about 2-3 hours; (debug logging is enabled)

Dec  2 18:37:33 pixelserv-tls[12612]: SSL_accept error:0 status:3
Dec  2 18:37:33 pixelserv-tls[12612]: SSL_accept error:0 status:3
Dec  2 18:37:33 pixelserv-tls[12612]: SSL_accept error:0 status:3
Dec  2 18:37:33 pixelserv-tls[12612]: SSL_accept error:0 status:3
Dec  2 18:37:33 pixelserv-tls[12612]: SSL_accept error:0 status:3
Dec  2 18:37:33 pixelserv-tls[12612]: SSL_accept error:0 status:3
Dec  2 18:37:33 pixelserv-tls[12612]: SSL_accept error:0 status:3
Dec  2 18:37:33 pixelserv-tls[12612]: SSL_accept error:0 status:3
Dec  2 18:37:38 pixelserv-tls[12612]: recv() ECONNRESET: Success
Dec  2 18:37:38 pixelserv-tls[12612]: Exit recv loop socket:10 rv:0 errno:0 wait_cnt:120 num_req:2
Dec  2 18:37:38 pixelserv-tls[12612]: shutdown() socket in thread or child process reported error: Transport endpoint is not connected
Dec  2 18:38:13 pixelserv-tls[12612]: recv() ECONNRESET: Success
Dec  2 18:38:13 pixelserv-tls[12612]: Exit recv loop socket:10 rv:0 errno:0 wait_cnt:120 num_req:2
Dec  2 18:38:13 pixelserv-tls[12612]: shutdown() socket in thread or child process reported error: Transport endpoint is not connected
Dec  2 18:38:17 pixelserv-tls[12612]: recv() ECONNRESET: Success
Dec  2 18:38:17 pixelserv-tls[12612]: Exit recv loop socket:10 rv:0 errno:0 wait_cnt:120 num_req:2
Dec  2 18:38:17 pixelserv-tls[12612]: shutdown() socket in thread or child process reported error: Transport endpoint is not connected

There are hundreds of them. Is it something to worry about?

 

[kvic]

@pattiri

Perfectly normal. E.g. each "SSL_accept error:0 status:3" is corresponding to +1 in counter 'slh' on servstats page.

If you want to log access URL/POST contents, use '-l 4' as suggested by 8. How do I capture access log to file? in FAQ.

'-l 5' means for DEBUG as suggested on servstats page. People shall not use it in general 'cos lot of "noise".

 

[quant88]

After installing v2.0.0-rc1, I was checking stats on 192.168.1.2/servstats.

Initially, it seems unresponsive on my windows PC. For some reasons, the IP was redirected by some malware to the below location

Tracing route to -sso.anbtr.com [192.168.1.2]
over a maximum of 30 hops:

1 2 ms 2 ms 2 ms -sso.anbtr.com [192.168.1.2]

Trace complete.

This is further confirmed - I am able to view the html stats on my android devices.
No luck with Malwarebytes, it is unable to find any malware on the windows PC..

 

[quant88]

After installing v2.0.0-rc1, I was checking stats on 192.168.1.2/servstats.

Initially, it seems unresponsive on my windows PC. For some reasons, the IP was redirected by some malware to the below location

Tracing route to -sso.anbtr.com [192.168.1.2]
over a maximum of 30 hops:

1 2 ms 2 ms 2 ms -sso.anbtr.com [192.168.1.2]

Trace complete.

This is further confirmed - I am able to view the html stats on my android devices.
No luck with Malwarebytes, it is unable to find any malware on the windows PC..

Cool.. Zemana AntiMalware found the culprit.

 

[quant88]

Cool.. Zemana AntiMalware found the culprit.

Tracert to 192.168.1.2 still direct me to sso.anbtr.com.

I was performing the tracert with Network Tools - Network Analysis.

traceroute to 192.168.1.2 (192.168.1.2), 30 hops max, 38 byte packets
1 -sso.anbtr.com (192.168.1.2) 0.135 ms 0.014 ms 0.009 ms

Can anyone help. Is this a concern?

 

[Correlor]

@quant88

In my setup doing a traceroute to my pixelserv-tls internal ip-address gives me the first domainname in my blacklist. So no malware but a result of using hostfiles to block domains.

 

[quant88]

@quant88

In my setup doing a traceroute to my pixelserv-tls internal ip-address gives me the first domainname in my blacklist. So no malware but a result of using hostfiles to block domains.

Thank you Correlor.
It is the first domainname in my blacklist too.

 

[thelonelycoder]

Tracert to 192.168.1.2 still direct me to sso.anbtr.com.

I was performing the tracert with Network Tools - Network Analysis.

traceroute to 192.168.1.2 (192.168.1.2), 30 hops max, 38 byte packets
1 -sso.anbtr.com (192.168.1.2) 0.135 ms 0.014 ms 0.009 ms

Can anyone help. Is this a concern?

This is how it works, if you ping that domain, omitting the dash*) it resolves to the pixelserv-tls IP.
The tracert / traceroute resolves 100% for all domains in the blacklist and the blocking file to the pixelserv-tls IP.

Make sure you are flushing the cache and only enter the IP/servstats, no http.

*) The dash is falsely added in two hosts files, the support.it-mate.co.uk/downloads/hosts.txt and hosts-file.net/emd.txt