• blah
  • blah

Note:  The following guide to Seti Queue is outdated.  There is a new version of Seti Queue that is available for download.   There is an updated guide for Seti Queue that former TLC member Geordie has written and is available for your perusal here.

Overview
As most people are well aware, the current stability of the Seti servers hasn’t always been the norm. Only just a couple of months ago, the servers were plagued with all sorts of issues, from bad memory modules and fault drives, to ‘routine’ maintenance that stretched from the foreseen two hours to several days in duration. All of these events took the servers offline, so that the average user was no longer able to get anything accomplished. Needless to say, it’s not a good event when it happens. After all, a good many of us have turned off the heat, so that we could save up enough money to buy more computer hardware. And when the Seti servers go down, the CPUs become idle, and the temperatures plummet. As you can see, a potentially life threatening situation for those of us in colder climes.

Enter Seti Queue. SetiQ was written by Reneris to help keep server outages from resulting in fatalities. Not only that, but it has a whole bunch of other useful features as well! The most obvious of these is it’s ability for you to keep on crunching even during extended server outages, which can be a huge advantage. Think of it this way- you’re able to crank out 10 units per day, and the Seti server crashes for 3 days. While most others (except those running caching programs) will become idle (and possibly frostbitten), your machines will continue to work away. You’ll get 30 work units done in those 3 days, while everyone else sits still. In effect, it’s like getting 30 free work units! And you stay warm to boot. 

Another great potential advantage is the connectivity end of SetiQ. For those who have a LAN environment, you can set the machine running SetiQ as a proxy server on all the other machines (we’ll go into more detail later on). When those clients finish, they’ll connect to the SetiQ server, upload and download a work unit, and continue crunching. For those of us with a ‘always on’ internet connection, such as DSL or cable, it’s not such a big deal. However, for those who still have dialup connections, the advantages can be significant. After all, dialup connections don’t always go through- and the Seti client will wait an hour before it tries again. Which means that you lose an hour of processing time. And frankly, with a small fleet of machines (say 5 or more), who wants to pay for all those modems? I’m depressed that I have to supply video cards for all of mine, seeing how I only have monitors for half of them. I’d hate to have to buy them all modems too! And for others, who have funny setups (one way cable modems come to mind), autoconnection isn’t an option, and this is a lot easier than writing scripts or batch files to cycle through your units.  Another advantage to running SetiQ is that your computer will officially be classified as a server. Not that it really matters, but everybody knows that most power hungry geeks long to have the ability to say that they have a server at home. J 

So, what exactly does SetiQ do? As I stated, it acts as a buffer between your clients and the Seti server. It will connect to the Seti server either manually, or at a time you specify, upload results, and download new work units for processing. The amount of units downloaded will vary depending on how many you can finish in one day, as well as how many days you’d like to buffer for, up to a max of 5. This is constantly recalculated and adjusted on the fly, so you don’t need to make any configuration changes if you add or remove a machine from the network. In addition to its basic use as a buffer, it can also log results for all the work units you’ve processed. It can also perform a job task when it completes with connecting to the server, such as writing a file to the desktop to let you know that it’s sent and received new units. However, I won’t be going over these last two abilities at all, since they aren’t really of any concern, except in certain situations outside the cares of the average user.

Setup
Setup of the SetiQ client is very easy. The entire configuration of the client is done with a simple text file, which comes with the client. All you have to do is edit a few things to make it meet your needs.  Here’s the settings you need to worry about, and where to set them.

 Work_unit_adjust         - set to the number of clients you have running. If you have 3 machines, set it to 3.

Sync_hour                     -set to whatever time you’d like to connect to the server automatically.  Don’t set between 5-7PM CST, you might miss zAmboni’s  stats update!

Autoconnect                  -if you want to connect manually, set this to no

Startupconnect              -set to no if you don’t want to connect automatically when the program starts

Queue_depth                 -set to 5

Response_delay            -set to 5 if you are using Setilog. Don’t forget to remove the # in front of the line

For most people, the rest of the settings don’t need to be touched. Also note that there is an explanation of each item in the file itself, which goes over the settings.

The other thing that will need to be configured is the Seti clients. For those with a fixed IP on their machine, you simply set the clients to use that IP address as the proxy server, with the port of 5512. So instead of starting Seti with “seti.exe”, your shortcut would use “seti.exe –proxy 192.168.0.1:5512”.  And with that, you’re all set to go!

Note that there are some additional settings that can be adjusted for the GUI. However, since I’m writing this at work (scary- I have more free time here than elsewhere!), and I have no ‘net access (boo!), I dunno no nothing about no GUI. Anyways, why are you running the GUI? It’s too slow.

Use
The first time you start SetiQ, you’ll notice that it doesn’t really do anything. It just sits there and states ‘Waiting for Seti client to send initial request.’  This is because SetiQ needs your user information. The easiest way to jump-start everything is to copy your Seti directory, and delete the workunit.sah file. Then start the client with the proxy setting, and it will connect to SetiQ and request a unit. SetiQ will then have all the info it needs, and can connect to the Seti server and download work units for your crunching enjoyment.

 There’s a handy tool that’s been included with SetiQ to send commands to the buffer, and that is SQ. SQ is a simple program that which is run from the command line, and only has 2 commands that you’ll really need to know. The first is the IP of the server. If you run it locally, use local. If it’s on another machine, use it’s IP address. The other command you need to know is connect, which is used to force SetiQ to connect manually. The IP (or local) must always be specified, so that SQ knows where to request info from. If you use ‘sq IP’, (substitute your IP address or ‘local’) the program will return the stats of the SetiQ program, including how many results are queued, how many work units are queued, and how many you finish per day, on average. If you use ‘sq IP connect’, the client will attempt to connect to the Seti server to send and receive more units.

The only thing left to be done now is to take care of the SetiQ window, which shows client connection info. I myself find it to be a pain in the butt, since it gets in my way. So you can either set it up as a service using srvany.exe (for those running NT or 2000), or hiding it using a program like Trayer. One advantage of using Trayer is that you can pull the window back up if you need to- although you generally can get any info you need using SQ.

Conclusion
As you can see, the SetiQ program has some nice merits to it. It can keep you from freezing to death in the winter, give you ‘free’ work units when the Seti server crashes, and can get rid of your server connection issues. And once you have it setup, you can just forget about it- it will take care of everything for you. As an added side bonus, it uses no CPU time, it’s memory footprint is very small, and it will work fine with clients from all OS types. So give the program a try, you honestly can’t go wrong.