Messages

46

Users / Re: MD recommendations

« on: January 14, 2014, 03:37:50 am »

Dual channel RAM kits would work best, so look for 2 x 1G, 2 x 2G, etc kits, depending on your budget.  Those little NUC's will handle up to 16G of RAM! 

Some folks are actually running home VMware ESX clusters on them..  Just Google "NUC ESX Server" for a few interesting home lab builds...

/Mike

47

Installation issues / Re: PXE boot MD

« on: January 14, 2014, 03:27:48 am »

Hi B34N!

Something isn't quite right then...  SSH into your core, and run the following commands:

ls -la /tftpboot
ls -la /tftpboot/default
ls -la /tftpboot/pxelinux.cfg

cat /etc/dhcp/dhcpd.conf

cat /var/cache/bind/db.linuxmce.local

Paste the output of each, and we'll see what the current state is...

HTH!

/Mike

48

Installation issues / Re: PXE boot MD

« on: January 13, 2014, 02:46:44 pm »

I have an old Pentium 4 laptop, 512M of RAM, that I was net-booting off of a highly modified Diskless MD image which I'm using as a workstation in my workshop.  I was using the local disk as swap, to offset the limited RAM.  That was working fine off of 10.04, so I think you'll probably be fine net booting it.

I did an in-place upgrade to 12.04, which I'll be documenting over time.  I'm sick right now, so I don't feel like doing it at the moment... The diskless image I was using caused me issues in the core upgrade (no real support for Diskless Workstations, only MD's) so I had to delete it.  If you're doing a new install of 12.04, things will likely go just as described in the Wiki.  Once I got over all the upgrade hurdles, 12.04 on the core has been running pretty well for me (Disclaimer: I'm not running the Diskless MD's yet; production is still on another network using Minimyth). 

HTH!

/Mike

49

Developers / DNS inconsistencies, input sought.

« on: January 09, 2014, 04:10:59 am »

Hi Folks!

Being a sysadmin type, I'm now trying to work my way through various sysadmin type Trac issues.  I just patched a NIS issue (Thanks Phenigma for committing it!), and now I'm currently working on a DNS related one.  I'm doing the relevant patches against 10.04, and will diff against 12.04 as well (I'm currently running 10.04, and plan to do an in-place upgrade).

http://svn.linuxmce.org/trac/ticket/1769

Some of this is easy to fix, but some needs some input.  Three scripts actually backup and manage aspects of resolv.conf on the core (trunk/src):
BootScripts/Network_DNS.sh
BootScripts/Network_Parameters.sh
BootScripts/Network_Setup.sh

Would it not be better/easier to maintain if only one script managed that resolv.conf file?  I may be missing out on why it is the way it is, so I thought I'd put it out there for those more knowledgable about LMCE internals to comment...

Thanks for your time!

/Mike

50

Developers / Re: New Template ISY994i uploaded

« on: December 14, 2013, 04:13:22 pm »

Chris,

That was a lingering bug I had fixed in the last update I'd committed.   I'd refactored the code at one point, and forgot to rename that function call which is sitting in the event handler in a different code file.

This is where you need to go through the web admin, Advanced -> SQLCVS -> Update and then wait a while.  When it says success, you can then reload the router and check the logs.  Triggering some lights should generate the status update events, and you'll find your errors will disappear.

If they don't clear up, let me know...

HTH!

/Mike

51

Developers / Re: New Template ISY994i uploaded

« on: December 14, 2013, 01:17:39 am »

edited: Basically, GSD _SUCKS_ when you need to maintain more than one socket. Full stop.
-Thom

Thom,

I can understand your position.  Practically speaking, with the mysql dependency satisfied, the driver works pretty well.  The ruby/GSD driver does make for a good prototype platform, to finish the feature set. It's taken me the better part of a year to get it to this point, and that is with some ruby knowledge. I had to teach myself some more advanced ruby, XML parsing, plus the ISY SDK.  If I had to learn C++ simultaneously, it'd be another year before I got anything out... 

I figure if I can get enough done and kitted out, someone more knowledgable than myself could port it to C++.

As a GSD, it's not perfect, but it does work well...

HTH!

/Mike

52

Developers / Re: New Template ISY994i uploaded

« on: December 14, 2013, 01:07:40 am »

Mike,

cat /var/log/pluto/55_Generic_Serial_Device.log
http://pastebin.com/itjtqGvP

Did you install ruby gems or some other package when your wrote the template.

Chris

Chris,

The pastebin isn't there anymore...

I didn't use ruby gems.  I have the following ruby-mysql packages on my core:
libmysql-ruby
libmysql-ruby1.8

apt-get install libmysql-ruby1.8 libmysql-ruby

would satisfy the mysql require.

HTH!

/Mike

53

Developers / Re: New Template ISY994i uploaded

« on: December 13, 2013, 06:56:11 pm »

Why does this driver require mysql?

Thom,

I'm querying the pluto database to look up the IP address of the device; I couldn't figure out a way to find it via device data.  If you can point me in the direction of that, I'll be able to remove the mysql requirement.  I thought the Ruby mysql drivers were installed as a LMCE dependency (which is why I was using them), but it could be that they are on my system due to some other products I'd installed.

The reader's digest version is that the conn_ method is used with the ISY's subscription channel (which is a one-way feed of events from the ISY to the "client", in this case the LMCE driver).  Commands have to be sent asynchronously as a separate REST call, and I have a function to do that.  The function hits the IP address, which I'm presently scraping from the DB, after first parsing the pluto.conf file to find the connection information.  The relevant DB scrape code is in the 355_Process_Initialize, and the REST command function is in 373_Private_Methods.  I can't establish the conn_ subscription until I've executed some REST calls to connect to the ISY and gather Child Device and other config data.  This was all done to make things work "automagically".

Chris,

Please stand by... based on Thom's assessment, your system is missing the ruby-mysql packages.  I suspect we'll have a way forward after Thom thwacks me upside the head... 

HTH!

/Mike

54

Developers / Re: New Template ISY994i uploaded

« on: December 13, 2013, 01:28:41 pm »

Hi Chris!

Man you're fast!  Been waiting for this for a while, eh?   

My first thought is that you need to go through the web admin, Advanced -> SQLCVS -> Update and then wait a while.  When it says success, you can either reload the router, or reboot the core to have all the auto detection stuff happen via dhcp.  It should detect the ISY, then create the device as a child of the core.  You'll need to change your ISY credentials to match via the web admin, and reload the router.  After that, things should just work as per the wiki page.

If you're still having problems after that, can you pastebin or post the log files here as attachments?  I'll need the <dev#>_Generic _Serial_Device.log and <dev#>_ISY-994i-GSD.log.  Both are located in /var/log/pluto, and can be found using

ls <dev#>_*

Mine are 189_Generic_Serial_Device.log and 189_ISY-994i-GSD.log (as an example).

I'll look at them tonight if you're still having problems.  I think you just got bit by the misaligned code (it got attached to a different device in the SQLCVS number shuffles...

/Mike

55

Developers / Re: New Template ISY994i uploaded

« on: December 13, 2013, 04:28:20 am »

Hi Thom!

I'm guessing that no news is good news, and that I'm on the right track.  I've done another anonymous SQLCVS commit. Would you be able to review it and let me know if it looks OK?  Dec 12th @ 22:20 EST, against Trac #1999.

Thanks!

/Mike

56

Developers / Re: New Template ISY994i uploaded

« on: December 09, 2013, 05:10:38 am »

Hi Thom!

I modified the template first to re-associate it with Universal Devices (#1163).  That's what I committed early this morning.  Later, I noticed the code wasn't associated, so I started looking into the tables. 

I modified table 'InfraredGroup_Command.FK_InfraredGroup' for all affected rows to match the relevant row from 'infraredGroup.PK_InfraredGroup'.  The old value was 6143 (which matched the Google ChromeCast values), and I changed it to 6102 for the ISY.  I noted the behavior for LMCE creating new command groups, deleted the new (blank) rows, and changed the FK_InfraredGroup value for the pre-existing ones. 

Let me know how you want me to proceed...

Thanks again!

/Mike

57

Developers / Re: New Template ISY994i uploaded

« on: December 09, 2013, 03:36:12 am »

Thom,

I've noticed that the code got tied to the Google manufacturer ID, which meant it wasn't working with the driver the way it should.  I've figured out how to fix it locally by modifying the database tables directly.  I'm not sure if doing a SQLCVS commit now will make things better or worse, so I'll wait on your advice.

Thanks for your assistance with this!  It's greatly appreciated!

/Mike

58

Developers / Re: New Template ISY994i uploaded

« on: December 08, 2013, 06:23:22 am »

Thom,

I've just committed a fix for Manufacturer ID.  It got switched to"Google" in the SQLCVS number shuffle, and was corrected to "Universal Devices".  Fix committed on Dec. 8 @ 0:18 EST via anonymous SQLCVS.

Thanks for the assistance!

/Mike

59

Developers / New Template ISY994i uploaded

« on: December 07, 2013, 04:32:01 am »

ISY994i Template Release notes

Initial checkin, Trac ticket #1999 committed Dec. 6, 2013 at 22:20 EST.  Would someone be able to review and approve it, please?  Edit: Forgot to mention, it's a SQLCVS anonymous commit...

Wiki Page for driver: http://wiki.linuxmce.org/index.php/ISY994i

Working Features:

Insteon/X10 Devices and Insteon Scenes
Automatic device creation in LMCE
Automatic creation of Insteon Scenes as dimmable child devices
Generates status update events to LMCE
Control of devices and scenes using orbiters

Future work/To-Do list:
- finish up device status, config routines and logic
- integrate orbiter alerts
- add more events support
- add more device support and verbosity
- Insteon Keypad support (basically implemented using related devices)
- relate ISY folders to LMCE rooms (and logic applied
- Integrate ISY WeatherBug module with 1204's WeatherPlugin

Thanks!

/Mike

60

Installation issues / Re: Problems intstalling to VirtualBox

« on: December 02, 2013, 01:23:02 am »

Specialstuff,

20G is probably too small.  I run mine with a 50Gb LV for root, and seperate LV's for /home and /var/log, and you can see what I'm using.  LMCE is virtualized using KVM, with LVM's passed through as block devices.  /var/lib/mysql is an LV from an SSD, and the large log directory is for centralized syslogging, and verbose logging for dev/debugging work.

Code: [Select]

Filesystem            Size  Used Avail Use% Mounted on
rootfs                 50G   36G   12G  76% /
none                  787M  316K  787M   1% /dev
/dev/disk/by-uuid/86982b08-7e50-blah
                       50G   36G   12G  76% /
/dev/vdc1              20G  1.8G   17G  10% /var/log
/dev/vde1             296G  2.9G  278G   2% /home
/dev/vdb1              12G  2.7G  8.6G  25% /var/lib/mysql

You have to remember, each MD is almost another install itself, averaging 2.5 G apiece.

Hope that helps!

/Mike