Messages

5. I don't think the "new" squeezebox radio is branded as such, but it looks almost the same.

It's called the Smart Radio UE, and runs a different firmware for their streaming services.  A "downgrade-able" firmware is available to restore compatibility with the Logitech Media Server (aka Squeezebox Server).  More info can be found in the Squeezebox Radio forum http://forums.slimdevices.com/forumdisplay.php?32-Squeezebox-Radio

I thought about getting one, but decided to just go with Raspberry Pi's. That way, I can use them for Airplay and as Squeezeslaves, and add home automation to the mix using AgoControl (as sensor platforms as well as maybe running LED lights via OpenDMX).  Still playing with it at this stage.

You might want to look into SqueezePlug, a ready to roll distro for the RaspberryPi.  http://www.squeezeplug.eu

Hope that helps!

/Mike

Pigdog,

Jumbo frames don't generally work well in mixed environments.  Usually, you use jumbo frames on network segments that don't have to route anywhere (vMotion, FCoE, dedicated storage network, backup network, backend networks, etc).

What's happening is the jumbo frames are having to be broken down when routed through the core, re-transmitted, and the received traffic will have to be chunked together to make new jumbo frames to send back to the MD's (incurring significant latency).  Needless to say, performance will suck royally (especially on commodity hardware).

You may wish to reconsider your usage strategy for jumbo frames...

If you choose to keep using them, you may wish to investigate/verify that your MD's are using the Apt-proxy and a http caching proxy, and you'll likely need to do some serious Kernel network performance tuning...

/etc/sysctl.conf

Code Select Expand


net.core.rmem_max = 16777216
net.core.wmem_max = 16777216
net.core.rmem_default = 16777216
net.core.wmem_default = 16777216
net.core.optmem_max = 16777216
net.ipv4.tcp_rmem = 4096 87380 16777216
net.ipv4.tcp_wmem = 4096 87380 16777216

The above is a place to start, and can be used on both core and MD's (even without jumbo frames).  It sets 16M buffers for the network, presuming 1Gbe network adapters.  I run those settings on my network without issues.  The kernel defaults to 1M and 4M buffers, which are more suitable to 100M network adapters.

Here's another place to look for ideas...
https://www.ibm.com/developerworks/community/wikis/home?lang=en#!/wiki/Welcome%20to%20High%20Performance%20Computing%20%28HPC%29%20Central/page/Linux%20System%20Tuning%20Recommendations

Just keep in mind, you won't be able to avoid the fragmentation issues; you may just be able to mitigate them to a tolerable extent.  It looks like you like to play, so I figured I'd point you towards some places to start...

HTH!

/Mike

N.B.  The IBM wiki page above also has tuning parameters for 'ib', which are Infiniband adapters. Disregard unless you actually have an Infiniband fabric in your house...

Joerod,

Speaking as a server guy, your "free" server may end up costing you in the long run.  It's a circa 2001 Dual-processor Pentium 3, with 133MHz (not GHz) RAM, Ultra SCSI 3 drives, 100M ethernet, and PCI/PCI-X slots all in a 7 RU chassis.  No offense intended, but these days we call that a 13 year old space heater.  The amount of compute capability vs the amount of power being used is quite disproportionate compared to even inexpensive modern hardware.  To be perfectly honest, most modern smartphones have more horsepower...

I don't want to dissuade you from using it, but you may want to set your expectations appropriately... Also, keep in mind that if you start to use it for "production" purposes at home, you are running a higher level of risk.  Your likelihood of getting spare parts is slim, and those disks are around 13 years old.  In the enterprise space, 3-5 years is considered old for disks, and 5-7 is considered old for server hardware.  New hardware uses less power and has on-average double the performance of two years previous.

Just wanted to set your expectations relative to your hardware's capabilities...

Hope that helps!

/Mike

Code Select Expand


#!/bin/bash
kill -KILL  `ps ax|grep Proxy_Orbiter|grep log|cut -b1-5`


Works a treat for me.

Sweeeeet!  Worked like a treat for me too!  That restarts them properly; now to figure out what's breaking them in the first place...

Thanks Posde!  Much appreciated!

/Mike

Hi Folks!

Not to hijack the thread; is the web orbiter code also used for the Proxy Orbiters (like for Roaming Orb)?  I have problems with those just dying, and having to reboot the core to get them to work.  Happens in both 10.04 and 12.04.

Now back to your regularly scheduled discussion thread... 

/Mike

Hi Phenigma!

Another question, did the mysql-ruby dependencies ever get added to the device template? 

Not yet; I plan on adding it in the next round of updates.  I'm working on integrating the ISY's WeatherBug module with Thom's Weather plugin, and updating some hardware support.

Thanks for following up on this!

/Mike

Chris,

How's this been working out for you?  I got a PM from someone having trouble, so just want to confirm no other problems out there...  I'm going on the assumption no news is good news...

Is all your Insteon hardware supported by my driver?  I only have lighting devices, so I haven't added support for motions or sensors, but I could if I have someone willing to be the test lab...

Cheers!

/Mike

I use MythTV 0.25 on my Debian host, and 5 MiniMyth diskless systems booted on my "external" network.  I'm using Schedules direct for the listing data, and have a mixture of Analog cable and Rogers Digital STB (PVR-150's on slave backend, PVR-1600 on host) as well as Digital OTA (HDHR-3-US on it's own network segment and NIC).

That's production for now, as I work on migrating to LMCE at some point in the (hopefully) near future.

I want to use mythtv but don't consider it usable within lmce.  I long for schedulesdirect and myth scheduling... it's SO much better than my satellite providers interface and capabilities!

Phenigma,  I'm curious as to why you don't consider MythTV usable within LMCE.  Cutting over to it was going to be one of the activities I was going to work on in the next couple of months, so I'm interested in knowing what I'm in store for... 

Thanks!

/Mike

MKBrown - sounds as if you have some idea there but please liase with possy before applying anything. The original problem that I noted above has definitely been resolved by his latest snapshot but there may still be some work to do. When I get a moment I'll read what you've written and try to understand it all.

Jamo,

I don't have commit rights to SVN, and to be quite honest, I don't want commit rights.  I'll plug away fixing stuff I'm capable of, usually SysAdmin type stuff 'cause that's what I do, do up patches and stuff them into Trac.  Someone more knowledgable than I about LMCE internals can take 'em, leave 'em, tweak 'em, whatever.  I'll just do what I can do, and let others do what they're good at.

Hope that clarifies things!

/Mike

Hi folks!

I've been looking into the DNS issues as well as the management of resolv.conf.  In 12.04, /etc/resolv.conf is now managed dynamically by the glibc resolver and resolvconf.  This is so that network mangler can dynamically manage Network configuration (and thus DNS lookups) on say a laptop that roams onto different Wi-Fi networks.  Historically, LMCE has been more like a server in trying to manage DNS settings statically, and now the two are conflicting to some extent.

I just finished working up a patch for Network_Setup.sh to add DNS settings into the /etc/network/interfaces file for the internal network, and when LMCE is running, the bind forwarders file handles outgoing DNS.  I haven't quite figured out exactly what to do about resolv.conf management, and at what stage to transfer the info obtained from Kubuntu's boot-up network config in resolv.conf over to the bind forwarders file.

Once set up properly, LMCE resolv.conf will look like this:

Code Select Expand

# Dynamic resolv.conf(5) file for glibc resolver(3) generated by resolvconf(8)
#     DO NOT EDIT THIS FILE BY HAND -- YOUR CHANGES WILL BE OVERWRITTEN
nameserver 127.0.0.1
search linuxmce.local


And the interfaces file will look like this:

Code Select Expand

# --- Internal NIC ---
iface eth1 inet static
        address 192.168.80.1
        netmask 255.255.255.0
        pre-up sysctl -q -e -w  net.ipv6.conf.eth1.disable_ipv6=0
        # DNS Settings for Internal Net
        dns-nameservers 192.168.80.1
        dns-search LinuxMCE.local


There won't be any DNS info in the external NIC (unless I add code for it; the challenge is again when to transfer it). External DNS is handled by the servers listed in /etc/bind/named.conf.forwarders, which is also setup in Network_Bind.sh, which is called from Network_Setup.sh. 

To make it interesting, the following scripts all backup or touch resolv.conf:

Code Select Expand

Network_DisplaySettings.sh:     local netExtDNS1=$(cat /etc/resolv.conf | grep "^nameserver" | head -1 | awk '{print $2}')
Network_DisplaySettings.sh:     local netExtDNS2=$(cat /etc/resolv.conf | grep "^nameserver" | tail -1 | awk '{print $2}')
Network_DNS.sh:if [ ! -e /etc/resolv.conf.pbackup ] ;then
Network_DNS.sh: cp /etc/resolv.conf /etc/resolv.conf.pbackup
Network_DNS.sh:if ! BlacklistConfFiles '/etc/resolv.conf' && ! BlacklistConfFiles '/etc/bind/named.conf.forwarders' ;then
Network_DNS.sh:$(grep nameserver /etc/resolv.conf | grep -v '#' | sed 's/nameserver//g; s/ *//g; s/^.*$/\t&;/')
Network_Parameters.sh:  if [ ! -e /etc/resolv.conf.pbackup ] ;then
Network_Parameters.sh:          cp /etc/resolv.conf /etc/resolv.conf.pbackup
Network_Parameters.sh:  if ! BlacklistConfFiles '/etc/resolv.conf' ;then
Network_Parameters.sh:          grep nameserver /etc/resolv.conf | grep -v '#' | sed 's/nameserver//g; s/ *//g' | tr '\n' ',' | sed 's/,$//'
Network_Setup.sh:# Empty resolv.conf files. Will be filled with static entries, or autofilled by dhcp
Network_Setup.sh:if [ ! -e /etc/resolv.conf.pbackup ] ;then
Network_Setup.sh:       cp /etc/resolv.conf /etc/resolv.conf.pbackup
Network_Setup.sh:: >/etc/resolv.conf
Network_Setup.sh:               if ! BlacklistConfFiles '/etc/resolv.conf' ;then
Network_Setup.sh:                               echo "nameserver $i" >>/etc/resolv.conf
Network_Setup.sh:               if ! BlacklistConfFiles '/etc/resolv.conf' ;then
Network_Setup.sh:                               echo "nameserver $i" >>/etc/resolv.conf


And the diskless setup scripts copy the core's resolv.conf to the moon directory, which won't work too well when the core is running properly, as resolv.conf will be set up to use localhost to hit bind on the same system. The diskless code will need a version of my patch to add the DNS settings to the interfaces file, or write the core's internal IP to resolv.conf.

I'll post my interface DNS patch into Trac shortly, so at least one issue gets addressed while we figure out the resolv.conf one.

EDIT:  added my patch to Jamo's ticket #2073.

Hope that helps!

/Mike

Ok, if you're running VirtualBox, you don't need 32-bit libraries on the host.  VirtualBox provides hardware virtualization, so all you need to do is create the VM "container" with at least a 50GB  "virtual hard drive", attach the LMCE ISO to it as a CDROM, and configure the boot order to boot from CDROM first.

You may wish to read through the VirtualBox docs at https://www.virtualbox.org/manual/ch01.html

You can use VBox as a test bed, but I wouldn't recommend it for "production" use, at least not with LMCE.  Unless you really know your way around virtualization products and configuration, it's best to just install LMCE one bare metal.

HTH!

/Mike

Thom,

Will do.  I try to understand how things work and why, before I dive in (which means things take a while ;-). I'll look into the resolv.conf stuff shortly.  I've submitted patches for 12.04 (which I'm now running) against ticket 1769 which should fix the local DNS updates from dhcpd. 

HTH!

/Mike

What's the host & hypervisor?  I do virtualization for a living, and short of architecture or endian issues, I have yet to see a 64-bit CPU that won't run 32-bit code (with the correct libraries of course); and that's talking Sparc, Power, and x86.

I run KVM on a Phenom II host with a 64-bit Kernel, 32-bit user land, and LMCE 32-bit running as a VM.  Before anyone thinks I'm crazy (hmmm, maybe? ;-). The host OS started life as a 32-bit Debian 3 install on an over-clocked Pentium 233 MHz, and has been through a number of in-place OS and hardware upgrades over the years, now currently Debian 7.1.  Multi-arch may let me finally go all 64-bit, or I may bite the bullet and re-install.  So much to do, so little time...

FYI, there is no current 64-bit LMCE builds...

HTH!

/Mike

mkbrown,

is there a reason why you specify sh as the shell to run Diskless_CreateTBZ.sh instead of relying on the shebang line in the script itself?

Mostly to credit huh with the fix, and from working on RHEL and Debian where /bin/sh is normally linked to bash, but your point is valid; especially since Ubuntu is linking sh to dash now... 

/usr/pluto/bin/Diskless_CreateTBZ.sh

is the proper way to execute the script B34N.  Thanks to posde for reminding me not to "assume".  I'll put it down to being sick right now... 

HTH!

/Mike

B34N,

Your issue is definitely due to the Diskless_BuildDefaultImage.sh script, which is called by Diskless_CreateTBZ.sh, not having been run.

As Huh indicated, you'll need to

sh /usr/pluto/bin/Diskless_CreateTBZ.sh

That will prepare the diskless images, and based on the other pastes, everything should just work at that point.

Thanks for the assist, huh!

/Mike