Old 5th December 2011, 22:54   #1
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
Plethora of problems...

hi,

im having a strange problem with the admin.cgi page in the latest linux 32bit build (running on centOS), the crux of it is, it only outputs half an html document, this seems to happen more often when the listenner count is higher.

clutching at straws my guesses are some sort of nasty character being thrown into the mix, some sort of maximum execution time being applied to the .cgi file (however it loads for 1-2seconds max not 10+ like you would expect) or something totally bizar i havnt thought of, have you ever seen/heard of anything like this, last line of document below:

<td align="center"><font class="sd">AppleCoreMedia/1.0.0.9A405 (iPhone; U; CPU OS 5_0_1 like Mac OS X; en_us)</font></td><td align="center"><font class="sd">24 minutes 15 seconds </font></td><td align="center"><font class="sd"><a id="st" href="/admin.cgi?sid=1&mode=kickdst&kickdst=-1281351780">Kick</a></font></td><td align="center"><font class="sd"><a id="st" href="/admin.cgi?sid=1&mode=bandst&bandst=178.98.189.129&banmsk=255&kickdst=-1281351780">Ban</a></font></td><td align="center"><font class="sd"><a id="st" href="/admin.cgi?sid=1&mode=bandst&bandst=178.98.189.129&banmsk=0&kickdst=-1281351780">Ban </a></font></td><td align="center"><font class="sd">

and it just stops, browser receives no more data...
Fibrewire is offline   Reply With Quote
Old 5th December 2011, 23:04   #2
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
hi im having a problem with only half the admin.cgi page is being transmitted,

running latest sc_serv build on a 32bit centos box

i could blather but this image sums it up

ive testing with FF same problem occurs, any ideas would be greatly appreciated
Attached Thumbnails
Click image for larger version

Name:	UPZ86.png
Views:	826
Size:	99.0 KB
ID:	49333  
Fibrewire is offline   Reply With Quote
Old 6th December 2011, 09:01   #3
DrO
 
Join Date: Sep 2003
Posts: 27,873
is it consistently happening and what number of clients are we talking about when you say 'listener counts are high' ?

it's not an issue i've seen / heard of happening before though i have changed some of the handling with that list since the build this thread relates to was released so it may already have been resolved (not sure in all honesty).

-daz
DrO is offline   Reply With Quote
Old 10th December 2011, 11:20   #4
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
Hi Dr0,

yes this is consistently happening,

My apologies I should of qualified, about 50+ client connections seem to cause this behaviour.

I downloaded the latest build from the website about 1 week ago , judging by the file name I im guessing you have not updated it since then sc_serv2_linux_07_31_2011.tar.gz

this is the last line of code the browser is receiveing:

<td align="center"><font class="sd"><a id="st" href="/admin.cgi?sid=1&mode=kickdst&kickdst=-1278195476">Kick</a></font></td><td align="center"><font class="sd">

(i can pastebin the entire source if it would help)

this behavour is repeatable in chrome/FF/IE/Opera/Safari

the server is running centOS 5.2
Fibrewire is offline   Reply With Quote
Old 10th December 2011, 11:58   #5
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
Interestingly

I built a new identical relay from scratch, this is repeatable on all relays but it always cuts off in the same place,

[between BAN IP and BAN SUBNET of the 11th entry] <-sratch that

it jossles back and forth between Kick IP, Ban IP, Ban Subnet & Reserve IP but it does always appear to the on the 11th entry so far

i hope this helps
Fibrewire is offline   Reply With Quote
Old 10th December 2011, 12:00   #6
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
just a thought,

could this have something to do with logging,

ive disabled all logging of any kind by the relay imaginable, is it possible that all it can output in the set amount of time the snapshot lasts for is 11 entries?
Fibrewire is offline   Reply With Quote
Old 14th December 2011, 13:52   #7
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
Plethora of problems...

so im running radiopayback nowadays

we have a multi server setup, with 1 private server and multiple public relays for listenners to connect to

all running the latest version of DNAS, the private server is running on MS Server 2008, the public relays are all running on centOS 5.2

the number of weird/stupid problems im experiencing are rediculeous... they do all revolve around the admin view on the server however.

  1. Primary the tail log does nothing of any use...

    msg:<***> Logger startup
    msg:<***> version 2.0.0.29
    msg:<***>
    msg:[SHOUTcast] DNAS/posix(linux x86) v2.0.0.29 (Jul 31 2011) starting up...
    msg:[MAIN] PID: 32219
    msg:[MAIN] Loaded config from /root/SHOUTcast/sc_serv_tune.conf
    msg:[MAIN] Calculated CPU count is 4
    msg:[MAIN] Starting 4 network threads
    msg:[MICROSERVER] Listening for connection on port 8000
    msg:Error opening port 8000 because Could not bind to :8000 because Address already in use
    msg:[MAIN] Runner shutdown
    msg:<***> Logger shutdown

    Then nothing... I appreciate that there is a message there that says "Error opening port 8000 because Could not bind to :8000 because Address already in use" however i call BS on that since the relay is working...
    ---------------------------------------------------------------------------------------
  2. The Listenner view...

    It cuts off after the first 11-20 entries for no good reason, its as if the cgi script has timed out (it hasn't it runs for way under 1 second), just stopping mid word/link/code
    ---------------------------------------------------------------------------------------
  3. Innacurate Listenner Count...

    im certain that it isnt detected disconnects from listenners, reporting a vast number of connections even after i have manually kicked everyone on the relay, this is locking up the relays when they reach capacity even though there really isnt anybody tuned in.

config:

Quote:
adaptivebuffersize=1
admincssfile=v2
adminpassword=[Redacted]
allowpublicrelay=0
allowrelay=0
autodumpsourcetime=0
autodumpusers=0
backupfile=
banfile=/SHOUTcast/control/sc_serv.ban
bufferhardlimit=16777216
buffertype=0
configrewrite=1
cpucount=0
destip=
faviconfile=
faviconfilemimetype=image/x-icon
fixedbuffersize=1048576
flashpolicyfile=/SHOUTcast/crossdomain.xml
hidestats=0
httpstyledebug=1
introfile=
listenertime=0
log=1
logclients=1
logfile=/SHOUTcast/logs/sc_serv.log
maxheaderlinecount=100
maxheaderlinesize=2048
maxhttpredirects=5
maxspecialfilesize=30000000
maxuser=49
metainterval=8192
microserverdebug=0
namelookups=0
password=[Redacted]
portbase=8000
publicserver=never
relayconnectretries=9999
relaydebug=0
relayreconnecttime=10
relayshoutcastdebug=0
relayuvoxdebug=0
requirestreamconfigs=1
ripfile=/SHOUTcast/control/sc_serv.rip
riponly=0
robotstxtfile=
rtmpclientdebug=0
savebanlistonexit=1
saveriplistonexit=1
screenlog=1
shoutcast1clientdebug=0
shoutcast2clientdebug=0
shoutcastsourcedebug=0
songhistory=10
specialfiletmpdir=
srcip=
statsdebug=0
streamauthhash_1=[Redacted]
streamdatadebug=0
streamid_1=1
streampath_1=/relay
streamrelayurl_1=[Redacted]
threadrunnerdebug=0
titleformat=
unique=$
urlformat=
uvox2sourcedebug=0
uvoxcipherkey=foobar
w3cenable=1
w3clog=/SHOUTcast/logs/w3c.log
webclientdebug=0
yp1debug=0
yp2=1
yp2debug=0
ypaddr=yp.shoutcast.com
ypmaxretries=10
ypminreportinterval=10
yport=80
yppath=/yp2
ypport=80
ypreportinterval=300
yptimeout=60
if anyone else has experience/resolved any of these issues, there is a compatability issue with centOS or there is error with my config file please let me know as we have a major event fast approaching.
Fibrewire is offline   Reply With Quote
Old 14th December 2011, 14:06   #8
DrO
 
Join Date: Sep 2003
Posts: 27,873
i've merged all of your posts into the same thread and i will reply to the points raised shortly (had meant to reply to the earlier ones in the DNAS thread but it's either reply or code things at the moment so it looks like coding is now out of the window whilst i properly reply).

-daz
DrO is offline   Reply With Quote
Old 14th December 2011, 14:12   #9
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
Dr0 thats fine bro i appreciate the situation

just glad to hear im not talking @ a wall

if you like i will gladly PM you the relay admin panel passwords and modify the config for debugging if it helps you
Fibrewire is offline   Reply With Quote
Old 14th December 2011, 15:00   #10
DrO
 
Join Date: Sep 2003
Posts: 27,873
with the admin page failure, i think that's more likely coming from an issue in the processing of the client details since it should not be just stopping and would point possibly at the data coming out not being liked by the browser which is causing the connection to be stopped. or it's in the DNAS itself (i know i've changed some handling on that page with the client list since the public build which may / may not have any effect). additionally not having logging enabled is not going to cause such issues since the admin page output is not related to the logging options in anyway.

Quote:
Originally Posted by Fibrewire View Post
Primary the tail log does nothing of any use...

msg:<***> Logger startup
msg:<***> version 2.0.0.29
msg:<***>
msg:[SHOUTcast] DNAS/posix(linux x86) v2.0.0.29 (Jul 31 2011) starting up...
msg:[MAIN] PID: 32219
msg:[MAIN] Loaded config from /root/SHOUTcast/sc_serv_tune.conf
msg:[MAIN] Calculated CPU count is 4
msg:[MAIN] Starting 4 network threads
msg:[MICROSERVER] Listening for connection on port 8000
msg:Error opening port 8000 because Could not bind to :8000 because Address already in use
msg:[MAIN] Runner shutdown
msg:<***> Logger shutdown

Then nothing... I appreciate that there is a message there that says "Error opening port 8000 because Could not bind to :8000 because Address already in use" however i call BS on that since the relay is working...
The only thing i can think off is that the log _can_ be overwritten if a second instance of the DNAS is started with the same details as the previously started DNAS instance, which would also match up with what is being seen.

Only solution for that is to restart the original DNAS or try to do a forced rotate of the log file via the server admin summary page or sending a SIGHUP to the running instance.


Quote:
Originally Posted by Fibrewire View Post
The Listenner view...

It cuts off after the first 11-20 entries for no good reason, its as if the cgi script has timed out (it hasn't it runs for way under 1 second), just stopping mid word/link/code
as mentioned at the start of my post, i think it's more likely the page is outputting something which isn't right and i'd have expected to have heard more about this than just you being the only one who has reported such an issue.

for checking that out, i would suggest having a look at the 'listeners' page shown on the server summary page (admin.cgi or admin.cgi?sid=0) and see what the xml output shows (as that shouldn't fail to keep outputting and may give a better reason why it's not working / make sure that there is nothing dodgy about the clients).


Quote:
Originally Posted by Fibrewire View Post
Innacurate Listenner Count...

im certain that it isnt detected disconnects from listenners, reporting a vast number of connections even after i have manually kicked everyone on the relay, this is locking up the relays when they reach capacity even though there really isnt anybody tuned in.
i've never seen it act weird in the handling of this so i'm firstly not sure where you're getting the values to determine that is the case. kicking is not an instance action and can take a while for it to be fully applied as the data and connection is dropped but it's also determined by how the client connection. as client connections _may_ potentially remain at times up to the timeout depending on how they have or haven't closed the connection with the DNAS. if something is causing the connection to remain open as is covered by the timeouts (which are up to 30 seconds) then i cannot see anything being wrong.


nothing jumps out from the config file though as everything including defaults has been dumped i could be missing something though i don't think so - will be glad when i finish the changes which makes configrewrite=1 only output non-default values for such things (configrewrite=2 will be the equivalent of the existing behaviour).

-daz
DrO is offline   Reply With Quote
Old 14th December 2011, 16:51   #11
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
Dr0 i just had the cunning idea of opening the webpage in notepad++ as modern browsers always want webpages to work and will fill in the blanks where as notepad++ wont.

hope this gives you some hint, currently all of my relays are displaying this behaviour so i have a good sample base

last line R1 (line 144)

HTML Code:
<td align="center"><font class="sd">Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.1; Trident/4.0; GTB7.2; SLCC2; .NET CLR 2.0.50727; .NET CLR 3.5.30729; .NET CLR
last line R2 (line 139)

HTML Code:
<td align="center"><font class="sd"><a id="st" href="/admin.cgi?sid=1&mode=bandst&bandst=140.124.131.136&banmsk=0&kickdst=140269236">Ban </a></font></
last line R3 (line 409)

HTML Code:
<td align="center"><font class="sd"><a id="st" href
judging by where the lines cut off, 3 different places, mid tag etc etc, i would say something is severely f000ked, and that it is not likely to be the only user based input which would be the "User Agent" field, which surely you would be parsing anyway to stop injected code.
Fibrewire is offline   Reply With Quote
Old 14th December 2011, 16:57   #12
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
Logs,

with regards to logs i believe you were correct, the first line of my now working tail log reads:

2011-12-14 18:29:31 I msg:[MAIN] Rotated log file

i didnt quite grasp your explanation before, could you please go into more detail specifically on how i can resolve this problem

on a seperate note the second line of my now working tail log file reads:

2011-12-14 18:29:31 E msg:virtual void AOL_logger::consoleLogger_element::log(const AOL_logger::message&) Error writing to console - write issue

now i am the first to admit that i suck when it comes to linux, and i believe that the insane way i am starting the shoutcast server on boot means that there may well be no console as such for it to write to, in which case i can simple change screenlog off, does that legit/correct?
Fibrewire is offline   Reply With Quote
Old 14th December 2011, 17:20   #13
DrO
 
Join Date: Sep 2003
Posts: 27,873
Quote:
Originally Posted by Fibrewire View Post
Dr0 i just had the cunning idea of opening the webpage in notepad++ as modern browsers always want webpages to work and will fill in the blanks where as notepad++ wont.

hope this gives you some hint, currently all of my relays are displaying this behaviour so i have a good sample base
the option i mentioned from the server summary page would be far easier to work with and actually validate against rather than trying to work against the incomplete webpage output which isn't giving me much of anything to go against.

Quote:
Originally Posted by Fibrewire View Post
judging by where the lines cut off, 3 different places, mid tag etc etc, i would say something is severely f000ked, and that it is not likely to be the only user based input which would be the "User Agent" field, which surely you would be parsing anyway to stop injected code.
or it could be something i've already fixed, i just don't know since the public build is massively out-dated compared to the internal builds. i'd really need the xml output as mentioned so i can test against that.


Quote:
Originally Posted by Fibrewire View Post
i didnt quite grasp your explanation before, could you please go into more detail specifically on how i can resolve this problem
basically don't try to re-run the same DNAS+config file instance when one is already running. that's all there is to it.


Quote:
Originally Posted by Fibrewire View Post
on a seperate note the second line of my now working tail log file reads:

2011-12-14 18:29:31 E msg:virtual void AOL_logger::consoleLogger_element::log(const AOL_logger::message&) Error writing to console - write issue

now i am the first to admit that i suck when it comes to linux, and i believe that the insane way i am starting the shoutcast server on boot means that there may well be no console as such for it to write to, in which case i can simple change screenlog off, does that legit/correct?
it usually happens when the shell the DNAS is running in has gone away (either due to it timing out or closing due to other factors). if you don't need to see the console output then screenlog=0 is fine and will prevent the chance of those errors appearing.

-daz
DrO is offline   Reply With Quote
Old 14th December 2011, 18:14   #14
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
admin.cgi?sid=1&mode=viewxml&page=3

looks solid to me, is there another page you'd like me to try?

HTML Code:
<?xml version="1.0" encoding="UTF-8" standalone="yes" ?>
<SHOUTCASTSERVER>
  <LISTENERS>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>Mozilla/5.0 (Ubuntu; X11; Linux x86_64; rv:8.0) Gecko/20100101 Firefox/8.0</USERAGENT>
      <CONNECTTIME>2391</CONNECTTIME>
      <UID>0x85bcb74</UID>
    </LISTENER>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>Mozilla/5.0 (Windows NT 6.1; WOW64; rv:9.0) Gecko/20100101 Firefox/9.0</USERAGENT>
      <CONNECTTIME>1324</CONNECTTIME>
      <UID>0x85be8d4</UID>
    </LISTENER>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>Mozilla/5.0 (Windows NT 6.1; WOW64; rv:8.0) Gecko/20100101 Firefox/8.0</USERAGENT>
      <CONNECTTIME>1337</CONNECTTIME>
      <UID>0x85cc664</UID>
    </LISTENER>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.1; Trident/4.0; GTB7.2; SLCC2; .NET CLR 2.0.50727; .NET CLR 3.5.30729; .NET CLR 3.0.30729; Media Center PC 6.0; .NET4.0C; .NET4.0E; CMDTDF; BRI/1; InfoPath.3; Tablet PC 2.0; BRI/2)</USERAGENT>
      <CONNECTTIME>3932</CONNECTTIME>
      <UID>0xb3a01aec</UID>
    </LISTENER>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>Opera/9.80 (Windows NT 6.1; U; ru) Presto/2.10.229 Version/11.60</USERAGENT>
      <CONNECTTIME>4060</CONNECTTIME>
      <UID>0xb3a01c5c</UID>
    </LISTENER>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>Mozilla/5.0 (Windows NT 6.1; WOW64; rv:8.0) Gecko/20100101 Firefox/8.0</USERAGENT>
      <CONNECTTIME>4076</CONNECTTIME>
      <UID>0xb3a044f4</UID>
    </LISTENER>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>Mozilla/5.0 (Windows NT 6.0; WOW64; rv:8.0.1) Gecko/20100101 Firefox/8.0.1</USERAGENT>
      <CONNECTTIME>4073</CONNECTTIME>
      <UID>0xb3b012ac</UID>
    </LISTENER>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>Mozilla/5.0 (Windows NT 6.1; rv:8.0.1) Gecko/20100101 Firefox/8.0.1</USERAGENT>
      <CONNECTTIME>2806</CONNECTTIME>
      <UID>0xb3b0697c</UID>
    </LISTENER>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>XMPlay/3.6.0.1</USERAGENT>
      <CONNECTTIME>1817</CONNECTTIME>
      <UID>0xb3b07144</UID>
    </LISTENER>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.1; Trident/4.0; GTB7.2; SLCC2; .NET CLR 2.0.50727; .NET CLR 3.5.30729; .NET CLR 3.0.30729; Media Center PC 6.0; .NET4.0C; .NET4.0E; CMDTDF; BRI/1; InfoPath.3; Tablet PC 2.0; BRI/2)</USERAGENT>
      <CONNECTTIME>3927</CONNECTTIME>
      <UID>0xb3b08f9c</UID>
    </LISTENER>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>Mozilla/5.0 (Windows NT 5.1) AppleWebKit/535.2 (KHTML, like Gecko) Chrome/15.0.874.121 Safari/535.2</USERAGENT>
      <CONNECTTIME>469</CONNECTTIME>
      <UID>0xb3b0e4cc</UID>
    </LISTENER>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/535.2 (KHTML, like Gecko) Ubuntu/11.10 Chromium/15.0.874.106 Chrome/15.0.874.106 Safari/535.2</USERAGENT>
      <CONNECTTIME>2045</CONNECTTIME>
      <UID>0xb3c00fd4</UID>
    </LISTENER>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.1; WOW64; Trident/4.0; SLCC2; .NET CLR 2.0.50727; .NET CLR 3.5.30729; .NET CLR 3.0.30729; Media Center PC 6.0; MDDC; .NET4.0C; Creative AutoUpdate v1.40.04)</USERAGENT>
      <CONNECTTIME>4075</CONNECTTIME>
      <UID>0xb3c04b3c</UID>
    </LISTENER>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>Mozilla/5.0 (Macintosh; Intel Mac OS X 10_6_8) AppleWebKit/535.7 (KHTML, like Gecko) Chrome/16.0.912.63 Safari/535.7</USERAGENT>
      <CONNECTTIME>1341</CONNECTTIME>
      <UID>0xb3c07354</UID>
    </LISTENER>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/535.7 (KHTML, like Gecko) Chrome/16.0.912.63 Safari/535.7</USERAGENT>
      <CONNECTTIME>467</CONNECTTIME>
      <UID>0xb3c07ba4</UID>
    </LISTENER>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/535.2 (KHTML, like Gecko) Chrome/15.0.874.121 Safari/535.2</USERAGENT>
      <CONNECTTIME>2475</CONNECTTIME>
      <UID>0xb3c0e634</UID>
    </LISTENER>
    <LISTENER>
      <HOSTNAME></HOSTNAME>
      <USERAGENT>Opera/9.80 (Windows NT 5.1; U; en) Presto/2.10.229 Version/11.60</USERAGENT>
      <CONNECTTIME>1340</CONNECTTIME>
      <UID>0xb3c0e78c</UID>
    </LISTENER>
  </LISTENERS>
</SHOUTCASTSERVER>
Fibrewire is offline   Reply With Quote
Old 14th December 2011, 19:35   #15
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
2 listenners showing, all other stats show 41, including the XML pages
Attached Thumbnails
Click image for larger version

Name:	QPPyo.png
Views:	171
Size:	15.3 KB
ID:	49351  
Fibrewire is offline   Reply With Quote
Old 15th December 2011, 10:24   #16
DrO
 
Join Date: Sep 2003
Posts: 27,873
i'll send you a link to a test build to try out at some point today as from the quick look i've been able to do, i think it's something i've already fixed from 'Fixed the 'Unique Listeners' total on the administrator summary page not always reporting the correct number' (taken from the 60 line changelog for the next build) plus it'll be easier for any additional testing since that was one of the earliest changes i made after the public release and we're almost 6months later than that.

and i'll add that i would trust what is generated by the xml output and shown in the summary totals against the table since as that is failing then that will not be complete (and is created from the same information as the xml and summary totals).

-daz
DrO is offline   Reply With Quote
Old 15th December 2011, 15:01   #17
DrO
 
Join Date: Sep 2003
Posts: 27,873
pm now sent so now i wait and see if that fixes it or not...

-daz
DrO is offline   Reply With Quote
Old 16th December 2011, 09:41   #18
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
Thanks Dr0 i will test this within an hour or two and get back to you
Fibrewire is offline   Reply With Quote
Old 18th December 2011, 20:47   #19
DrO
 
Join Date: Sep 2003
Posts: 27,873
any luck with the test build? (thought 2.5 days was long enough to wait before bumping)

-daz
DrO is offline   Reply With Quote
Old 19th December 2011, 11:20   #20
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
Hey Dr0,

its hard to say since im watching for it not breaking, however, i haven't noticed any of the problems mentioned previously so far, what is the ETA on this build being available for windows/linux 32/64 bit, as we have a variety of OS's and a wider deployment would giving a larger test set.

one request though if i may, would be some way of highlighting non unique listenners, its all well and good it telling you how many duplicate connections there are, but it is long/difficult to filter them out by hand, this is likely something that can be done with a bit of PHP admittedly, but it seems like something sensible enough to build in
Fibrewire is offline   Reply With Quote
Old 19th December 2011, 11:53   #21
DrO
 
Join Date: Sep 2003
Posts: 27,873
Quote:
Originally Posted by Fibrewire View Post
its hard to say since im watching for it not breaking, however, i haven't noticed any of the problems mentioned previously so far
that's really all i needed to know if the test build was working as expected or if the issues with the listeners list was still present.

Quote:
Originally Posted by Fibrewire View Post
what is the ETA on this build being available for windows/linux 32/64 bit, as we have a variety of OS's and a wider deployment would giving a larger test set.
there is no current eta, currently it's when it's ready which is not the case at the moment.

Quote:
Originally Posted by Fibrewire View Post
one request though if i may, would be some way of highlighting non unique listenners, its all well and good it telling you how many duplicate connections there are, but it is long/difficult to filter them out by hand, this is likely something that can be done with a bit of PHP admittedly, but it seems like something sensible enough to build in
the only obvious thing jump to mind on how to show it which wouldn't be a pain to implement / make it too confusing would be to group the details as per the mock up i've attached. only thing is that it then makes the list a bit funky looking since it's currently sorted by the client connection length though i guess it depends on how many non-unique are connected as to how confusing it might be.

-daz
Attached Thumbnails
Click image for larger version

Name:	unique_merge_mock.png
Views:	1535
Size:	16.7 KB
ID:	49365  
DrO is offline   Reply With Quote
Old 19th December 2011, 12:40   #22
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
i think that looks kinda cool, a far simpler way of doing it could just to be leave the list as is but sort by IP rather than connection time, that way all the duplicates would be side by side rather than having to actually group them together, or even just using a little CSS to highlight the duplicate entry line a different color

our relay number 4 is currently showing: Stream Listeners: 100 (74 unique)

like hell am i going to sit here and filter the 26 duplicates one by one (would be ok if it was 1 asshole connected 26 times but that isnt the case here), and since the list view on this relay cuts off after the... 41st entry, it becomes even more disheartening.

the downside however is thats 26 listenners worth of bandwidth wasted / second, and 26 people who could be enjoying some music who cannot
Fibrewire is offline   Reply With Quote
Old 19th December 2011, 12:57   #23
DrO
 
Join Date: Sep 2003
Posts: 27,873
i've just implemented what i put in the mock though have changed it a bit so there's just one kick option (which then will apply it to all clients on that ip which i think is how it should be anyway based on the column's action).

not sure on changing what things are sorted by though it looks like due to the additional processing i'm having to do it is now basing it on the client IP and then connection duration - not sure if that's going to be taken as a good thing or not seeing as it's been sorted by the client duration going back through all prior v2 and v1 builds...


with the uniqueness, i see fully where you're coming from though it could be they are legit unique listeners but how they're accessing the stream means they end up with the same IP via a proxy, corporate network, etc. though hopefully the grouping will make that a bit easier to work out what is going on.

-daz
DrO is offline   Reply With Quote
Old 19th December 2011, 15:49   #24
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
well if its grouped by IP then there's no reason to change it to sort by IP

that was just my suggestion for an alternate system that would give the same result

do you take donations for this? we get a lot of donations to the station and id be more than happy to filter some of that on to the development of a more robust system to run the station on.
Fibrewire is offline   Reply With Quote
Old 19th December 2011, 16:00   #25
DrO
 
Join Date: Sep 2003
Posts: 27,873
i went back to the prior method in the end so it's sorting by longest to shortest connection time (which is caught be the first IP in a group), and then sorting by the same within any IP groups so it looks much like before but with the addition of grouping (don't want to scare people too much, heh). as sorting by the IP groups as happened from my initial code implementation didn't look right.

-daz
DrO is offline   Reply With Quote
Old 19th December 2011, 16:27   #26
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
fair enough its still a viable solution to the problem, donations? (nevermind just seen your signature)
Fibrewire is offline   Reply With Quote
Old 21st December 2011, 12:02   #27
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
bad news,

ive now seen first hand the issues with the DNAS as described previously on the new build on 3/5 of our relays

1. innacurate listenners count (in my opinion it looks/feels like it isnt reporting disconnects, so the listenner count just climes and climes until eventually maxuser is hit)

2. listenners page only shows the first X entries then cuts off

I don't mind sending you and admin password for 1 of the relays so you can witness this if you like.
Fibrewire is offline   Reply With Quote
Old 21st December 2011, 12:21   #28
DrO
 
Join Date: Sep 2003
Posts: 27,873
k, pm or email me the details (address is on my site in my signature) and will try to have a look today.

-daz
DrO is offline   Reply With Quote
Old 21st December 2011, 12:52   #29
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
ill wait until it does it mate

(i did a force restart on each relay which did solve the problem)

but now ill have to wait a few hours for them to get "gunked" up again
Fibrewire is offline   Reply With Quote
Old 23rd December 2011, 13:48   #30
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
heres a different one for you:

http://r5.radiopayback.com:8000/crossdomain.xml

404 not found, fair enough...

msg:[MICROSERVER] Flash policy file server not enabled

?
Fibrewire is offline   Reply With Quote
Old 23rd December 2011, 13:51   #31
Power2All
Junior Member
 
Join Date: Dec 2011
Posts: 1
Hi, I am the one running r5.
The paths are direct to the file.
Also, the bans file in /control/ folder dissappear too after the server shutted down.
Is that normal ?

[Edit]
Bans file stays when there is data in it, so that looks normal.
Power2All is offline   Reply With Quote
Old 23rd December 2011, 14:01   #32
DrO
 
Join Date: Sep 2003
Posts: 27,873
Fibrewire: crossdomain.xml should work if you've specified a file to use via 'flashpolicyfile' or that there is a crossdomain.xml in the same folder as where the DNAS is running from. the 'Flash policy file server not enabled' is due to using an internal build which is not finished as that feature is not complete (and has since been disabled in the builds).

Power2All: the DNAS will remove the files when there is nothing needing to be saved in them.

-daz
DrO is offline   Reply With Quote
Old 23rd December 2011, 14:12   #33
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
Quote:
Originally Posted by DrO View Post
Fibrewire: crossdomain.xml should work if you've specified a file to use via 'flashpolicyfile' or that there is a crossdomain.xml in the same folder as where the DNAS is running from. the 'Flash policy file server not enabled' is due to using an internal build which is not finished as that feature is not complete (and has since been disabled in the builds).

Power2All: the DNAS will remove the files when there is nothing needing to be saved in them.

-daz
works ok here: http://r3.radiopayback.com:8000/crossdomain.xml
Fibrewire is offline   Reply With Quote
Old 23rd December 2011, 14:51   #34
DrO
 
Join Date: Sep 2003
Posts: 27,873
all i can say is to check the config file to make sure it's trying to load something which exists and that no errors are otherwise present in the log when trying to access things. otherwise there's really not too much else i'm able to do at the moment as i'm tight on available time (which means i don't think i'm going to be able to finishing trying to reproduce and fix the listener lists issue until either later part of next week or into january).

-daz
DrO is offline   Reply With Quote
Old 23rd December 2011, 15:38   #35
Fibrewire
Junior Member
 
Join Date: Dec 2011
Posts: 21
Thats fine buddy, just trying to give you all available information, enjoy XMAS & New Year
Fibrewire is offline   Reply With Quote
Reply
Go Back   Winamp & Shoutcast Forums > Shoutcast > Shoutcast Discussions

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump