Home of the Squeezebox™ & Transporter® network music players.
Page 3 of 3 FirstFirst 123
Results 21 to 25 of 25
  1. #21
    Senior Member
    Join Date
    May 2008
    Location
    Canada
    Posts
    5,061
    Reason why I asked about jumbo frames and MTU is because I had a sort of similar problem a few years ago with a NAS that was activating such jumbo frames when it should have not. But that was probably an old buggy stack. I know itĺs mainly a discussion with you and @bpa now but it seems to me that the only different variable in your system is your NUC so it feels like it has to be something related to its IP stack. Why a precise song? Maybe because due to the compression ratio, the classic starts to control flow at a different point of time that happens at the wrong moment during some TCP/IP process (hence I was also asking for window size values)

    Good luck
    LMS 7.7, 7.8 and 7.9 - 5xRadio, 3xBoom, 4xDuet, 1xTouch, 1 SB2. Sonos PLAY:3, PLAY:5, Marantz NR1603, JBL OnBeat, XBoxOne, XBMC, Foobar2000, ShairPortW, JRiver 21, 2xChromecast Audio, Chromecast v1 and v2, , Pi B3, B2, Pi B+, 2xPi A+, Odroid-C1, Odroid-C2, Cubie2, Yamaha WX-010, AppleTV 4, Airport Express, GGMM E5

  2. #22
    Junior Member
    Join Date
    Sep 2019
    Posts
    18
    So I wanted to tie this up.

    I stripped out all extra code I'd been messing with - waiting for STMh, using select, setting nonblocking, all the many things I tried - and moved my server to my linux laptop.

    Everything works fine. The buffer fullness climbs straight to 0.99 and hovers there; songs play through without a hitch. A blocking socket, writing 65536 bytes at a time, works perfectly well, exactly as I'd expect.

    Now what I have to do is figure out why a wired connection from a raspberry pi worked fine, and a wired connection from a laptop running Linux 4.15.0-55-generic, Mint, works fine, but an otherwise idle NUC running Linux 4.15.0-64-generic, Mint, with a wired connection to the same switch, stalls horribly (but somehow can manage other network traffic, like running a mumble server on occasion, without an issue.) I would not have believed it, and I will be playing with cables, switches and, probably, learning more about configuring TCP on linux than I ever wanted to. Or, I'll buy a $35 dollar pi to stream my music again.

    This is, frankly, creepy. I would never have considered Linux capable of stalling a socket so badly, especially when it has nothing else to do. I hope to mercy it's not some problem with the NUC hardware itself.

    Anyway, thanks for the support.

  3. #23
    Senior Member
    Join Date
    Oct 2005
    Location
    Ireland
    Posts
    17,864
    Quote Originally Posted by ScottAM View Post
    This is, frankly, creepy. I would never have considered Linux capable of stalling a socket so badly, especially when it has nothing else to do. I hope to mercy it's not some problem with the NUC hardware itself.
    Thanks for giving an update of where you are. I may sound like a cracked record but the benefit of a pcap recording is that you get to see the network interaction right down to physical level. I once had a device with a faulty socket which meant short packets were OK but long(i.e 1500 bytes) became occasionally corrupted - TCP covered up all the errors but problem was easily seen by looking at netstat numbers and pcap dumps. Unforunatety pcap recording may be one-sided as it is hard (not impossible) to setup an extra device to be a passive monitor of the network comms from the NUC.

    If the issue is with the NUC - you have the usual suspects - drivers, hardware and interconnections.

  4. #24
    Junior Member
    Join Date
    Sep 2019
    Posts
    18

    Solved

    The problem turned out to be weirder than I imagined, and I had imagined quite a lot.

    Having both the server's wired and wireless devices on caused the problem. The moment I shut the wireless off, everything got normal.

    Thanks to folk who listened to me whine and opine, and pushed me into looking at tcpdump - that was when I realized I had two active interfaces. I don't know why having both of them active matters, but it absolutely does.

    Anyway I have my server talking to a Squeezebox running 81, a squeezebox running 131, and a piCorePlayer, with no issues. Thanks again.

  5. #25
    Senior Member
    Join Date
    Oct 2005
    Location
    Ireland
    Posts
    17,864
    Good to know problem was sorted. I hate when weird problems are left hanging .
    Some issues with two net interfaces have come up before (mainly configuring LMS) but I can't remember any matching your symptoms.

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •