Self-Installer didn't work or, worse yet, currently fictional? This is the place to be. This guide will guide you through setting up CASJobs, armed with nothing but a mere zip file and an agile clicking finger.
CASJobs requires that the following be active before installation. Unless otherwise noted, these do not all have to be installed on the same machine.
A brief description of each component of this system follows.
These components may be run on any combination of machines, so long as they are on the same local network. For instance, one might have the website on one machine and BatchAdmin and the JobsService on another.
In the interest of brevity, well-documented database commands or procedures will often be glossed over. Books-Online for SQL server (BOL) is a good reference and installs by default to Microsoft SQL Server -> Documentation and Tutorials -> SQL Server Books Online.
You should have a file called CASJobs_install.zip. Upon extraction you should see, at the root level, a file 'cjemptybackup.bak' and another directory. That directory will be the root of the website; you should now copy that elsewhere. A good place for it is under IIS's root directory, generally at 'C:\Inetpub\wwwroot'. 'cjemptybackup.bak' is temporary and can go anywhere.
You should now create an IIS web application for CASJobs. This can be done by opening up IIS manager (under administrative tools) and then finding the directory called 'CASJobs' you just copied in the previous step. Right-click that directory, choose properties, then find where it says 'Application Name' and then click the 'Create' button on the same row. That's it; click ok and you're done. You can verify that this worked by opening a web browser pointing at the appropriate, local URL. By default, this is http://localhost/casjobs. Before continuing, you should make sure this page is loading correctly. If it's not loading, verify that
A few website files are not checked into cvs directly, but are instead checked in with the name of their organization appended to the file name. This is so fresh checkouts do not overwrite existing changes. Currently there two sets of templates, 'sdss' and 'af'. These are essentially themes. 'sdss' is a dark, night sky theme, while 'af' has a more earthy tone. Upon a new installation you should copy all files from one template to files without the extension. Below is a list of all the template files.
CASJobs requires SQL authentication for all databases used. If anything is set to exclusively use windows authentication, then it will not work. A database can be configured to use mixed-mode authentication by doing the following in Microsoft SQL Server Management Studio (SSMS) in SS2k5, or Enterprise Manager in older SS versions.
right-click server -> properties -> security -> Server authentication = 'SQL Server and Windows Authentication mode'
Each and every database server used by CASJobs must be set to use this form of authentication.
'BatchAdmin' is the name of the database CASJobs uses to manage many back-end tasks. To set it up, first use SSMS to restore the file cjemptybackup.bak to a database called 'BatchAdmin'. Some weirdness on using SSMS to restore a database from a file...
Tables not mentioned are only used internally, and should not be modifed by a person.
I had trouble getting this to work properly using just the SSMS GUI hooks. However, the follow SQL commands should work just fine. You should connect as an admin user and run these under the 'BatchAdmin' database, not master.
use [BatchAdmin] EXEC sp_adduser 'batch'
EXEC sp_addlogin 'batch', 'password', 'BatchAdmin'
use [BatchAdmin] EXEC sp_password NULL, 'abc@BC123', 'batch'
EXEC sp_change_users_login 'Update_One', 'batch', 'batch';
ConnectionStrings are a way of referring to a particular database on a particular machine. All databases in CASJobs are described with a connectionString at some point, somewhere. When setting up a database, you should keep track of the following information, since you'll probably have to enter it somewhere else later.
Below follows a connection string example...
Initial Catalog=name_of_db; Data Source=name_of_machine; User ID=batch; Password=pw;
It consists of '=' separated pairs, which are in turn separated by ';'s. Spacing is important, 'DataSource' is not the same as 'Data Source'.
It is expected that the database servers that make up the 'Data database[s]' and the 'Mydb database[s]' are all linked together. For example, the user that logins into a MyDB database should be able to query each data database from that MyDB connection. You can setup linked servers in SSMS or use the T-SQL command sp_addlinkedservers. You might find it easier to manage users if you setup all data and mydb databases to use the same one.
This is where the 'batch' user will be creating MyDBs (user databases). This server uses two SQL Server users, one for initially creating databases, and one for accessing them after that. This is to prevent users from such mischief as increasing the size of their own MyDB, dropping their own MyDB, etc.
Once you've setup the database users, you should update the MyDBHosts table in BatchAdmin. Below follows an explanation of the columns in MyDBHosts.
When users refer to a catalog in CASJobs, such as 'DR4', they are referring to a virtual catalog, not a specific machine. The 'Servers' table contains all the mappings from context names (dr4) to actual machines. A machine is usually determined by a queue, target pair. However, queue, target duplicates are allowed; in this case, queries sent to that queue, target will alternate, performing rudimentary load balancing. Each 'target' field (called 'context' elsewhere), should have at least two entries; a short or 'quick' version with a queue value of one, and a longer version, with a queue value greater than one. Even if a context has just one machine for both quick and long jobs, there should still be two entries for this context, which the same machine listed in each.
JobsService is a service that constantly runs in the background, picking up and executing asynchronous queries and output jobs. To run JobsService, it must first be installed with a program called 'installutil.exe' that comes installed with .Net. By default, this program is not included in the path. It's located wherever .Net installed itself, which, on this particular machine is...
C:\WINDOWS\Microsoft.NET\Framework\v2.0.50727\InstallUtil.exe
The JobsService executable is located under the CASJobs directory here:
\JobsService\bin\Release\JobsService.exe
To install this service, simply pass JobsService as a parameter, ie
installutil.exe Jobsservice.exe
If you don't have installutil.exe in your path, you're probably going to have to give an absolute path to one of those. You'll have to tell a name to call the service; remember this name, since this is how you'll find it later. To uninstall the jobsservice, simply do the same thing, and pass the \u parameter. You'll also have to give it the credentials for the user you want to run it as. This can be a pain, since this dialog always pops up in the background (it should pop up after you enter the service name) and makes it seems like the installer is 'stalled' if you don't see it. Also, the user you give it should have the domain in it. So if your domain is 'DOMAIN' you'll have to give it a login like user@domain, etc...
Once installed, the service may be stopped/started from the services management panel, found under Administrative Tools -> Services
You should see the service here, as the name you gave it earlier. To start it, right-click it and choose start. This service must be running for asynchronous queries and output jobs to function.
The 'Output' table determines the format in which MyDB table contents may be downloaded and where these resources will be stored. The backup should have all available types of output already in it. You will have to set at least the path and url of each output type to something appropriate for your setup. Below follows the column descriptions...
blah blah blah explain
blah blah css