Cricalix.Net

Yeah, I’ve let things lapse a bit more than I normally do. New job has me quite occupied, and I had a house guest for the past month. House guest has now gone home, and I’m back to being a lazy bachelor. Yes, this means I need to do the dishes in the sink. Perhaps tomorrow.

Work has thrown up an interesting project recently; I’ve had to learn a fair bit more about Asterisk, and some lingo for a totally different world than computing. The project has also given me more reason to knuckle down and get a distribution server up and running, using knowledge from the previous job. It’s certainly different working for a (slightly) larger company – in comparison to both the previous job and my work in academia post-degree. That said, I’m enjoying it thoroughly, and picking up more Windows administration skills than I will ever admit to knowing (like using VBScript + HTA + ODBC to make a 30 minute task take 1 minute).

On the EVE front, the corporation I’m in is back into Empire space, having grown tired of the antics of the alliance we were in. Three other corporations have ever so kindly declared war against us, so we’re having fun shooting them, not playing by their rules, and never talking to them in the chat channels. The ‘never talk’ one seems to really get their goat. Led a fleet operation on Saturday, and managed to lose no ships while destroying four of the enemy. Only four battleships though, we missed the heavy interdictor.

Also on the EVE front, I’m doing more work on the kill board software that we use to track our kills and losses. I wrote a value editor module for version 1.2 of the kill board, and have been keeping it compliant with 1.2, 1.3 and the latest 1.4 release. Also been contributing patches to fix other issues with the software, though mostly minor ones. Assuming no one reports problems with the value editor, I should be submitting a rather large patch to the KB folks to get the latest version included into 1.4 (they have an old version, and it confuses people!).

And just because I’m living in the UK: My, what strange weather for this time of year.  Unseasonably warm.  Frogs are spawning already.

Note to self, when testing backup software that needs a gigabit link to get full performance, use the gigabit card for the network connection, not the 100 Mbit/s card.  I’ve been doing speed tests to test the multiplexing of Yosemite’s Backup, and it was capping out at 600 – 650 MB/minute, even with 8 servers as clients.  Two days into the testing, I realise that I forgot to move the appropriate IP to the gigabit fibre link.  Did that just now, and the backup server is pulling data in at 1.1 GB/minute or more.

Today, I had the joy of trying to troubleshoot why server A couldn’t talk to server B on port 5432. Packet traces on server A showed that the far end (or something in the way) supposedly said RST, ACK to any connection in the app server pool once that connection timed out. Packet traces on the far end showed no connection at all, so I decided to use telnet to test the connection.

-bash: telnet: command not found


So, given that telnet didn’t exist, I tried perl. Yep, perl exists. IO::Socket::INET exists too.

Code time!

use IO::Socket::INET;
$socket = IO::Socket::INET->new(PeerAddr => '10.151.2.1', PeerPort=>'5432', Proto => 'tcp') or die $!;
$socket->close();

That proved that I was actually getting connectivity to server B, so we were able to chase other avenues to solve the problem (and did – Sun’s app server has 3 ways of verifying that connections in the pool are working, and only one of them actually tests the network connection).

Networking gear can be irritating some days.  Case in point, the Catalyst 5500 at work, with its WS-5410 line card and WS-G5484 GBICs.

Everything on Cisco’s site says that these devices work together (and they do), and that the G5484 GBICs are 62.5/50 micron adapters designed to talk to multi-mode fibre.

Yet the only cable that works (ie, that makes the CatOS say the port is connected) is a 9 micron core single-mode fibre.

It. Should. Not. Work.

But. It. Does.

Does. Not. Compute.

I even tried the cabling and GBICs in another switch.  Same result.

Argh.

In a set of round-up links, James McGovern wonders if I’m aware of the fact that CIFS is considered legacy for interacting with Microsoft domains. I wasn’t, but I am now :) Alfresco, luckily, supports more than just CIFS for enterprise authentication, and way more than CIFS for presenting the data to users – and I had already started going down the path of LDAP or Kerberos based authentication for users. We’re also not exposing the CIFS service ability of Alfresco to our users either for now – it’s webdav and web browser for the forseeable future.

With the arrival of the internal deployment VM server at work, I’ve started the process of narrowing down exactly how I’m going to be providing server rebuild and management abilities. Prime candidates right now are the Kickstart capability of RedHat Enterprise (and any derived distributions such as CentOS) for automated deployment of servers, and Reductive Lab’s Puppet for management. I’m well aware that there are other automated management tools that can do what Puppet does, but as one article I read said – using any of them is better than using none of them. So I’m using Puppet.

For those who don’t know, Puppet is a Ruby tool that has two halves – puppetmaster on the central management server, and puppetd on the remote clients. All authentication is done with centrally signed SSL certificates, and all the communications appear to be encrypted with those certificates.

continue reading…

As mentioned in a previous post, Alfresco got put on my stack of things to look at. We’ve decided to forget about integrating it into Liferay for the time being, and just use Alfresco as a standalone product (possibly with LDAP or NTLM authentication). So, I gave a brief demo to my boss and my director today, and it was somewhere just above blind leading the blind :)

Alfresco is some very nifty software. It operates at least three different ways to get data in and out of it – a standard web interface, WebDAV and CIFS. With a few clicks, I can add the ‘Versionable’ aspect to every document loaded into Alfresco, and this versioning works regardless of which interface you use. So I can create an HTML document using the TinyMCE editor, and it gets version 1.0. I can then WebDAV in, copy the file to my PC, edit it, copy it back, and it gets version 1.1. For giggles, I can browse to the network share, edit the file ‘on the server’, and it gets version 1.2. Unfortunately, only the web interface lets you set a version note to go with the version. There’s also a drag-and-drop target executable that allows you to check a file out and in – so only your changes apply.

The WebDAV interface works from Konqueror – I’m having name resolution issues with the CIFS interface for some reason. I’m pretty sure I could get FUSE to talk to Alfresco, giving me a local mounted directory with full access into the Alfresco server.

So, onwards and upwards. Alfresco looks to be a nice solution to our morass of scattered documentation.