sc trans disconnecting

[systrum]

not sure why but when i leave sc trans to broadcast for itself it disconnects

there seems top be no prob when i am monitoring it or connected to my own stream however when i leave it alone and come back after some time i received the disconnect log below....it seems to reconnect no prob tho

not sure why......any clues?


<09/18/03@18:07:55> [STREAM] Disconnecting from stream host [waiting 10s]
<09/18/03@18:08:05> [STREAM] Creating stream socket
<09/18/03@18:08:05> [STREAM] Resolving stream host
<09/18/03@18:08:05> [STREAM] Connected to host server
<09/18/03@18:08:06> [STREAM] Sending stream information
<09/18/03@18:08:07> [MAIN] Title Updated
<09/18/03@18:11:48> [DECODE] Opened ****** Versus Japan - Go Hawaii.mp3
<09/18/03@18:12:22> [MAIN] Title Updated
<09/18/03@18:15:00> [STREAM] Disconnecting from stream host [waiting 10s]
<09/18/03@18:15:10> [STREAM] Creating stream socket
<09/18/03@18:15:10> [STREAM] Resolving stream host
<09/18/03@18:15:10> [STREAM] Connected to host server
<09/18/03@18:15:11> [STREAM] Sending stream information
<09/18/03@18:15:13> [MAIN] Title Updated
<09/18/03@18:19:33> [STREAM] Disconnecting from stream host [waiting 10s]
<09/18/03@18:19:43> [STREAM] Creating stream socket
<09/18/03@18:19:43> [STREAM] Resolving stream host
<09/18/03@18:19:43> [STREAM] Connected to host server
<09/18/03@18:19:44> [STREAM] Sending stream information
<09/18/03@18:19:45> [MAIN] Title Updated

[FesterHead]

Is sc_trans and the DNAS on ther same machine?

Is the specified mp3 bad?

[systrum]

no ....the dns is provided by fast-serv

this file has played ok before and this problem has occured with other files on a few occasions.

[DJ AmPs]

Try setting up a local DNAS and see if you have the same issue -- this will help eliminate (or show) mp3 files as the culprit.

[FesterHead]

Don't forget to configure the fast-serv DNAS to relay the local one.

[DJ AmPs]

Yep, efsterheads right.

if everything stays connected on your local DNAS set up fast-serv to relay it. This is usually the best way to steram to a remote DNAS anyways -- relaying adds more 'cushion' to the connection.

[systrum]

obviously doing something wrong as my fast serv server is not relaying the local one

i have specified the fast serv server to relay port 8000 and my local ip
but it shows up as being down, but my local is connecting to shoutcast fine.

i haven't specified something in the sc serv conf , i know it.....but not sure what ?

[systrum]

i fixed it....if anyone else has the same problem then please visit fast-serv and check in the knowledge base under skipping and disconnects

[systrum]

ok...back to this....this solved the issue for a while but then the problem became worse.
i spoke with the guy at fast-serv an he recommended i go back to just broadcasting with sctrans. not exactly sure why but after about a period of 1 to 3 or more hours (always various) the stream disconnects for some time before reconnecting again. pain in the ass.

not exactly sure what is causing it tho. thought that my server (fast serv - they provide the extra bandwidth) was the problem so i ran a local server and it still occurs. i thought that it might be cuz of my changing ip but it happens too frequently...ip's don't change that often. the other thing was dhcp lease expiration
I have set my jaguar cache cleaner (a maintanence program) to renew my dhcp lease every 24 hours...would this solve the issue. my router is configured to renew the lease every 720 hours which is the max setting. not sure if renewing my lease would break my connection but it doesnt seem to. my isp is 'so net' have a router modem. figure that my config may need to be changed. how do u ensure that your broadcast stream never breaks?

any ideas on how i could fix this?

[Jay]

are you sure it's not the mp3?

[systrum]

well what would indicate that it is. i just says disconnecting.
i mean, i kind of tested the mp3s that were sus by putting them at the start of the playlist and then starting sctrans again. no probs so i doubt that is the issue. also , sctrans usually indicates if there is an mp3 issue by saying that it couldn't open it or something along those lines and then disconnecting and reconnecting

that only happend once so i deleated that file and it has never happened since

[systrum]

am currently running a local dnas without fast-serv to see if there are any probs
there is a prob in the log below

not sure what it means tho. does anyone else?

p_tch] yp.shoutcast.com touched!
<09/23/03@15:46:11> [yp_tch] yp.shoutcast.com touched!
<09/23/03@15:49:12> [yp_tch] yp.shoutcast.com touched!
<09/23/03@15:52:13> [yp_tch] yp.shoutcast.com touched!
<09/23/03@15:55:15> [yp_tch] yp.shoutcast.com touched!
<09/23/03@15:58:16> [yp_tch] yp.shoutcast.com touched!
<09/23/03@16:00:44> [dest: 192.168.0.4] starting stream (UID: 2)[L: 1]{A: iTunes/4.0 (Macintosh; N; PPC)}(P: 0)
<09/23/03@16:01:17> [yp_tch] yp.shoutcast.com touched!
<09/23/03@16:02:17> [dest: 192.168.0.4] connection closed (92 seconds) (UID: 2)[L: 0]{Bytes: 976000}(P: 0)
<09/23/03@16:04:19> [yp_tch] yp.shoutcast.com touched!
<09/23/03@16:07:22> [yp_tch] yp.shoutcast.com touched!
<09/23/03@16:10:23> [yp_tch] yp.shoutcast.com gave error (nak)
<09/23/03@16:10:23> [yp_tch] yp.shoutcast.com gave extended error (100)
<09/23/03@16:11:10> [active] 0 listeners (0 unique)

[Jay]

can you be more specific on what you are questioning? The yp error or the connection close?

[systrum]

i guess the disconnecting is the priority , but as there seems to be a scarcity of explaination regarding the nak (100) error {if u can explain}@i would appreciate any information as to why i could be getting this error while running a local dnas. most of the threads questioning this error seem to be unanswered. and the FAQ containing important info doesn't seem to mention it. am i wrong on this? is this error simply due to my changing ip?
i am running the local dnas to see if i get the same disconnecting problem as when i broadcast to the fast-serv server. so far, apart from the nak (100) error there are no problems or disconnects. this leads me to think that the problem is associated with fast-serv but i am not sure. wondering if anyone else has had the same disconnect issue or if anyone can help me to solve it.

if u can help answer both of these questions i would appreciate it.

[Jay]

the disconnecting is just someone disconnecting from your stream, not the source disconnecting.

Your stream looks fine from that log, the yp error given that it is successful most of the time would indicate to me that it is not a major problem. If you stop seeing yp touched successfully in your shoutcast server for an extended amount of time then worry about it.

[systrum]

KXRM
thanx for clarifying the nak (100) issue a little....

but these are two separate logs.....the one at the very top was from when i was broadcasting to fast-serv using sctrans only. the second log is with my local dnas.


this log shows sc trans disconnecting from the host. am i not correct???


<09/18/03@18:07:55> [STREAM] Disconnecting from stream host [waiting 10s]
<09/18/03@18:08:05> [STREAM] Creating stream socket
<09/18/03@18:08:05> [STREAM] Resolving stream host
<09/18/03@18:08:05> [STREAM] Connected to host serve


are you saying that the top log 'disconnects' are listeners disconnecting from my stream???

there is no listener ip or anything shown.
i think that this shows sc trans disconnecting and am not sure why this happens so often.

[systrum]

i preferably want to run sc trans with fast serv (for the added band width) but am having probs so... am running a local as a test and got the nak (100)

thought they may be related

what do u think?

[Jay]

no I was talking about the shoutcast log you posted. As I said I think you should examine the mp3's you are using, as the disconnects seem to coincide with song changes.

[DJ AmPs]

could be a CPU issue. I've seen that happen on slower PC's because when the song "pre-buffers" no stream is being sent. If pre-buffering takes too long then you get a drop in teh connection. Incorrectly coded mp3 files or otherwise corrupted can also cause this.

[systrum]

can anyone please help me out?? i really need some advice

with regard to KXRM's last reply....how can i examine more closely any problems that may exist with the mp3's? ... so far i have been deleting and replacing the mp3's in my mp3 directory. they are all encoded at 320kbps /44100khz.
some disconnects occur with the same mp3's but sometimes those mp3's play fine without an disconnection too. have been replacing the sus mp3s anyway .

still get disconnects within a 7 hour period.

with regard toDJ AmPs reply....am running the broadcast from a macintosh, 400 mhz PowerPC G4 (768 MB ram). There should be sufficient power to run the broadcast without any issues...i am not running anything else during the broadcast.

the final point that is really adding confusion is that my IP seems to change on occasion which i have been told should not occur for a dynamic ip if i am sending or receiving data. the ip seems to change occasonally on disconnect but at other times stays the same after a disconnect. not sure if the disconnect is due to the change of ip or is the disconnect is causing the IP to change (if thats possible)

[DJ AmPs]

actually 400mhz is barely cutting it. true a PPC theoretically has more power than a pentium but it depends on how well the software is compiled, and WHO KNOWS what the nullsoft crew was up to when they compiled it and for exactly what processor it's optimized for (if any).

My theories --
Each time a new track loads there's a peak in CPU usage and the buffer falls back slightly while the new track loads. Eventually you run out of buffer and it reconnects to fix the issue.

With the winamp dsp you will notice skipping when this happens as the buffer is "re-filled" with duplicate data to prevent buffering.

I am guessing that SC_TRANS doesn't use this retarded method of preventing buffering, especially since it involves extra code and SC_TRANS is quite minimal.

What i can suggest is to renice sc_trans which is the *nix equivilant of raising the CPU priority in windows.

type the following (use caps if the filename is SC_TRANS -- it's case sensitive)--
ps -ax | grep sc_trans

find the PID of the running process.

Now type the following --
renice -19 -p PID

where PID is the number you got from the output of first command. This will increase priority to almost the highest possible.

now issue 'top' and watch the sc_trans process -- make sure under the NI column the new priority of -19 is visible.

Let it run and see how it goes......

[systrum]

Thankyou for your reply and advice, muchly appreciated. i did what u said and ran the 'top' command after all that. could not see what you mentioned as the NI column. the process output is below. am not exactly sure if what i did was correct. i typed 'ps -ax | grep sc_trans' in the comand line and the final line showed a PID of 476. following that i typed
renice -19 -p476 and all seemed ok.

from what i have, is this correct?

Processes: 39 total, 3 running, 36 sleeping... 95 threads 20:25:46
Load Avg: 1.38, 1.71, 1.58 CPU usage: 92.0% user, 8.0% sys, 0.0% idle
SharedLibs: num = 90, resident = 21.9M code, 2.58M data, 8.27M LinkEdit
MemRegions: num = 2104, resident = 29.8M + 5.38M private, 45.0M shared
PhysMem: 53.9M wired, 64.7M active, 155M inactive, 273M used, 495M free
VM: 1.22G + 61.8M 6720(0) pageins, 0(0) pageouts

PID COMMAND %CPU TIME #TH #PRTS #MREGS RPRVT RSHRD RSIZE VSIZE
498 top 8.8% 0:12.21 1 14 17 216K 336K 512K 13.6M
496 sc_trans 88.9% 1:42.11 3 14 26 1.92M 988K 2.36M 5.16M
495 tcsh 0.0% 0:00.03 1 10 16 352K 600K 788K 5.73M
494 login 0.0% 0:00.92 1 12 33 248K 388K 576K 13.7M
493 Terminal 1.7% 0:37.01 5 85 191 4.18M 11.7M 14.2M 60.6M
437 lookupd 0.0% 0:00.59 2 33 46 336K 520K 864K 14.9M
400 Safari 0.0% 0:37.61 5 110 200 9.54M 12.4M 18.7M 85.0M
394 Finder 0.0% 0:04.37 1 74 125 2.59M 8.06M 7.63M 53.5M
393 SystemUISe 0.0% 0:02.72 1 141 120 1.40M 3.98M 3.73M 47.4M
392 Dock 0.0% 0:03.70 2 83 130 824K 8.14M 5.92M 47.6M
386 pbs 0.0% 0:01.73 2 28 27 376K 932K 1.29M 14.4M
380 DirectoryS 0.0% 0:00.92 3 52 144 736K 3.40M 3.27M 21.4M
364 loginwindo 0.0% 0:02.78 4 154 129 2.21M 8.33M 8.47M 54.9M
359 cupsd 0.0% 0:00.27 1 8 18 188K 480K 416K 14.9M
356 cron 0.0% 0:00.00 1 8 17 76K 336K 164K 13.8M
354 automount 0.0% 0:00.03 2 22 21 136K 428K 588K 14.7M
348 inetd 0.0% 0:00.00 1 8 13 56K 316K 112K 1.28M
344 dnsupdate 0.0% 0:00.16 2 25 23 296K 696K 1.05M 14.3M
335 nfsiod 0.0% 0:00.00 1 8 12 12K 332K 80K 1.27M
334 nfsiod 0.0% 0:00.00 1 8 12 12K 332K 80K 1.27M
333 nfsiod 0.0% 0:00.00 1 8 12 12K 332K 80K 1.27M
332 nfsiod 0.0% 0:00.00 1 8 12 12K 332K 80K 1.27M
319 coreservic 0.0% 0:01.36 3 62 68 1.22M 2.56M 3.52M 19.8M
296 netinfod 0.0% 0:00.56 1 7 26 128K 432K 296K 1.66M
283 crashrepor 0.0% 0:00.00 1 14 14 64K 308K 128K 13.3M
280 mDNSRespon 0.0% 0:00.13 2 30 23 660K 360K 944K 14.2M
270 syslogd 0.0% 0:00.08 1 7 13 80K 320K 164K 1.28M
203 autodiskmo 0.0% 0:00.51 1 37 20 160K 744K 368K 14.6M
182 Window Man 0.0% 1:27.69 2 162 228 4.00M 21.3M 24.6M 63.4M
173 ATSServer 0.0% 0:09.23 2 34 116 1.03M 4.87M 3.46M 56.2M
131 SecuritySe 0.0% 0:00.32 1 31 23 252K 956K 824K 14.9M
113 Audiophile 0.0% 0:00.08 1 16 21 172K 672K 852K 14.8M
101 configd 0.0% 0:05.90 3 108 130 696K 1.19M 1.63M 16.1M
74 dynamic_pa 0.0% 0:00.00 1 10 15 68K 312K 100K 1.29M
70 update 0.0% 0:00.12 1 7 13 52K 304K 104K 1.26M
51 kextd 0.0% 0:02.98 2 18 23 848K 664K 828K 15.3M
2 mach_init 0.0% 0:00.19 2 100 15 80K 332K 172K 1.80M
1 init 0.0% 0:00.05 1 9 12 88K 316K 272K 1.28M
[Process completed]0.0% 0:30.60 28 0 - - - 45.8M 491M

[systrum]

i am not so clued in with the processes that i have been recommended. sorry if i sound like an idiot.
Im guessing that i missed something along the way....

after the doing the renice process it showed

renice: [Process exited - exit code 1]

[DJ AmPs]

make sure there's no space between -p and 496 (or whatever PID you use)

according to your example you didnt use a space.

looks like your top reading doesnt include "niceness" but thats ok... try the renice thing again and add a space...

BUT sc_trans is using 88% of your CPU and your load average is >1 which means your system is struggling as I had previously guessed... =)

[systrum]

DJ AmPs, thanks for helping to locate what seems to be the problem.

i have tried doing this process again (both with and without caps) however it doesn't show a PID for sc_trans for some reason. may have to wait til the listeners drop and then shut down and restart sc_trans. may work then

will post again with the results

thanx again

[DJ AmPs]

if you type 'top' again you can see the pid real easy since sc_trans will always be near the top of top. You may hape to press P (capital) to sort by CPU usage.

maybe sc_trans aint runnin? =)

one more note -- you can only increase the priority of a PID (lowering NICE level) if you are logged in as root. I'm not sure how it works in OS-X as far as who you are logged in as.

[systrum]

no sc trans was running ...

i did this process again after logging out and logging in again. this time typing ps -ax | grep sc_trans gave me the PID and i did the renice thing again with the same outcome.

it showe...d renice: [Process exited - exit code 1]

i gather that this is not the desired out come is... or is it?

[systrum]

ok, sc trans has been running for a long period without disconnects which looks very promising (touch wood). I am not sure if what i did was correct (and if anyone can confirm this for me i would appreciate it) but it looks much better than before.

[DJ AmPs]

Process exited might not mean it didn't work... normally renice doesnt output anything so the OS might just be letting you know the process is exiting =)

hope it helped, and good luck!

[systrum]

the stream ran for over 24hours befor 2 disconnects within a period of 20 mins occurred which is a vast improvement...mind you i also manually configured my ip and stopped using DHCP which i suspected was also causing issues. see, my ip changes after a disconnect on occasion so i have been attributing the disconnect to the changing ip.

is it theoretically possible for my ip to change due to a disconnect...if so i could assume that it is either a CPU or mp3 issue.
is there anyone who has experienced this problem ?
seriously, if someone can help me resolve this then u are a god...

[DJ AmPs]

If you stop using DHCP, you may lose connectivity altogether sith some ISP.

But I really think your problem is CPU related, so other than renicing your sc_trans there's nothing you can do because sc_trans is not open source.

[systrum]

ok thanks, i really appreciate your help and advice. i do have a final question tho. am i required to renice each time i restart sc_trans? or is it a constant setting?

[DJ AmPs]

you are required to renice it each time...