Then the lightbulb turned on in my head! Why not just create the tablespaces in the Oracle Instance shell myself prior to running the installer and pointing it at the Oracle Instance. We are using Oracle 10G R2 RAC with 11G R1 clusterware on ASM storage. OMF (Oracle Managed Files) are enabled:

SQL> show parameter create_file
NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
db_create_file_dest                  string      +ASM_DBARCH_DATA

So after you execute Blackboard’s required grant step in SQLPlus as oracle:

@$ORACLE_HOME/rdbms/admin/rstrconn.sql;
GRANT create session, create table TO CONNECT;
commit;

…

Then execute this SQL to create the “base” set of Oracle Tablespaces for Blackboard:

CREATE SMALLFILE TABLESPACE "BB_BB60_STATS_DATA" LOGGING;
CREATE SMALLFILE TABLESPACE "BB_BB60_STATS_INDX" LOGGING;
CREATE SMALLFILE TABLESPACE "BBADMIN_DATA" LOGGING;
CREATE SMALLFILE TABLESPACE "BBADMIN_INDX" LOGGING;
CREATE SMALLFILE TABLESPACE "BB_BB60_DATA" LOGGING;
CREATE SMALLFILE TABLESPACE "BB_BB60_INDX" LOGGING; 

note that these SQLs do not cover the CMS tablespaces needed if you’re running the Blackboard (Xythos) CMS.

Here’s the SQL if you’re running the Blackboard (Xythos) CMS:

CREATE SMALLFILE TABLESPACE "CMS_DATA" LOGGING;
CREATE SMALLFILE TABLESPACE "CMS_INDX" LOGGING;
CREATE SMALLFILE TABLESPACE "CMS_FILES_USERS_DATA" LOGGING;
CREATE SMALLFILE TABLESPACE "CMS_FILES_USERS_INDX" LOGGING;
CREATE SMALLFILE TABLESPACE "CMS_FILES_COURSES_DATA" LOGGING;
CREATE SMALLFILE TABLESPACE "CMS_FILES_COURSES_INDX" LOGGING;
CREATE SMALLFILE TABLESPACE "CMS_FILES_ORGS_DATA" LOGGING;
CREATE SMALLFILE TABLESPACE "CMS_FILES_ORGS_INDX" LOGGING;
CREATE SMALLFILE TABLESPACE "CMS_FILES_INST_DATA" LOGGING;
CREATE SMALLFILE TABLESPACE "CMS_FILES_INST_INDX" LOGGING;
CREATE SMALLFILE TABLESPACE "CMS_FILES_LIBRARY_DATA" LOGGING;
CREATE SMALLFILE TABLESPACE "CMS_FILES_LIBRARY_INDX" LOGGING;

Note that this may not be an inclusive list for all installs, but for our work here with Blackboard 7.* through 9.0 SP1 HF2, this has been sufficient. If your installer still errors out after you’ve pre-created the tablespaces, you should investigate the log file that the installer creates and find the name of the tablespace that the installer is failing on. Then , using SQL like what’s above, create whatever tablespace you need.

I found out about all of this when we migrated from OCFS2-backed 10G R2 RAC without OMF (Oracle Managed Files) to ASM-backed 10G R2 with OMF on 11G R1 clusterware, but the solution is actually very simple if you look in the bb-installer.log that the installer creates when you run the jar like /usr/local/jdk/bin/java -jar bb-as-linux-9.0.505.15.jar. When I ran the installer with a /fakepath for the database install path, I saw the failing CREATE TABLESPACE statements referencing the /fakepath. I didn’t want BB to try to name files for the tablespaces, that’s why I’m running Oracle Managed Files. Once I pre-created the tablespaces, everything ran great and the install completed successfully. Detailed solution with example SQL after the break!

I have seen a lot of solutions come across the BBADMIN-L list out of ASU involving complicated solutions where one installs Blackboard 6,7,8 or 9 on non-ASM storage and uses RMAN to migrate the database. Another workaround involved creating a non-clustered database in a RAC environment on local storage and then moving that to the RAC clustered Oracle DB and ASM storage. Both of these solutions time consuming, fraught with places to make mistakes and hose the entire process, and generally a royal pain in the ass. I have a feeling Blackboard already knows this trick and if we’re lucky they’ll eventually publish their version of this article, if they have one.

I hope this helps all my fellow BBAdmins out just a bit. If you’ve got questions or this process blows up on you, please leave me a comment and we can try to figure it out! I have tested this on Oracle 10G R2, both RAC and non-RAC and with Blackboard 7.3.230.0, 8.0.375.0 and 9.0 SP1 HF2, all of which worked AOK.

Hive13 from Soapboxmedia.com on Vimeo.

The Security Center and You

On the test system (Coke) there is a new homepage module that everyone will see when they log in, it is called the security center. Please log in to coke (link only works on campus!) and see things for yourself. You’ll need an account on coke to do this, and you’ll login with your 6+2 and current LDAP/OneStop/etc. password. The Security Center was something I wrote to provide feedback about password related issues such as expiration and grace logins. If the user’s password is expires, but they still have grace logins remaining on their account, they will be able to login to Blackboard but it will consume a grace login and decrement the user’s grace logins available by one. By default, all UC accounts provide six grace logins, and therefor users have six logins before they are locked out and must change their password. I am investigating a way to display grace login and password expiration date in SysOp, but it’s not there yet.

Things to look for in SysOp

Examine Unique Ids Source when you lookup the user’s data. There is a known issue with accounts coming from “B – Grandfathered Beecher accounts”, those accounts do not exist in the LDAP tree that Blackboard authenticates against, and the caller should use their standard 6+2 login name with blackboard instead. This issue should be very rare. When Blackboard attempts to authenticate these users against LDAP, they will be treated in the same way as visitor and other non-LDAP accounts, meaning Blackboard will fall back to their RDBMS (database) password. If the user wishes to continue using the account, inform them they must login with the same password as they did before the conversion to LDAP since their account is being handled as a non-LDAP account.

Sysop showing my LDAP-authenticated account: if the password is expired, there will be a message showing it’s expired

Sysop showing showing a non-LDAP account

Here are a few scenarios that could occur when a user calls with password issues, beyond the standard bad password errors you are accustomed to.

First Scenario

User is unable to log in and receives message a message that their password is expired. The user’s password has expired and they have no grace logins remaining. The user must visit the UCit Password Self Service portal and reset their password, or contact the UC Help Desk for further assistance.

Second Scenario

User’s password is expired, but they were able login by consuming a grace login. Immediately advise the user to visit the UCit Password Self Service before they run out of grace logins. Nota bene: logging in to Password Self Service with an expired password requires at least one available grace login! If the user has just used their last grace login by logging in to Blackboard, and the Security Center displays a value of zero (0) for remaining grace logins, the user must contact the UCit Help Desk to reset their account.

Third scenario

User’s password will expired in less than the warning period (160 days in this case, but it is set to 7 days on the production system, this is for illustration only.) Advise the user to visit the UCit Password Self Service before their password expires.

Fourth Scenario

All is well, user’s password will expire within the configured seven-day warning period. I guess it’s not really a scenario at all.

Password Resets for accounts

Can be done through the UCit Password Self Service portal, or by having the user contact the UCit Help Desk.

Security Center and non-LDAP accounts

Security Center displays a message that the user’s password will not expire. Eventually this module will be hidden for non-LDAP authenticated user accounts. Of course, that “eventually” may be before December 22nd, but not at the time of the writing of this document.

Password Resets for non-LDAP accounts

I am currently doing the engineering before coding my own password reset utility for Blackboard to accommodate the fact that the built-in password reset functionality is not available when bb-config.properties is configured for LDAP authentication. This is very frustrating and I may give Blackboard a call to see if they can tell me how to force it on even with LDAP turned on. If this functionality is not restored, password resets for non-LDAP users will to be handled by the FTRC Blackboard Support Team via phone.

In Conclusion

If you run into issues not covered by this post, please feel free to come find me and ask. Once we solve the issue, we might even add it here. Consider this an experiment in work blogging. I’m looking forward to your comments on this post.

Via Hack a Day.

First thing to check is your signal strength, which is usually on a page called “Signal” for most motoLOLa cable modems, but I’ve seen it displayed on the main page too. There’s no hard and fast rule about where the number is located, but it’s almost guaranteed to be in dBmV, or decibel millivolts for those of us who aren’t engineers and managed to get laid in college. Chances are your signal will be a negative number, too. You want your signal to be better than -3dBmV or so. -4dBmV is probably fine too, but that’s about as low as I like to go. That’s fine, all the negative number means is that there is less signal at your modem than when it left the headend box at the cable company. If it’s not, there may be a line amplifier in your cable installation, which is common in apartment buildings and other shared structures.

If this number is very low, like -18dBmV (the worst I’ve ever seen), then you’ve got another problem – possibly bad cabling or too many splitters ahead of your modem. Most 2-way cable splitters will be a pair of -3.5dB outputs – meaning that each output is -3.5dB (a bit more than half) the strength than the input. This is cumulative, too. Meaning if you’ve got 10dB of signal coming into your house, and you split it twice with two 2-way splitters prior to your modem, you’re only going to have 3dB of signal left when you get to the modem. For the best connection, get the modem as close to the main cable drop as possible. Ideally, put a 2-way splitter in on the main drop, feed the cable modem with one output, and the rest of your systems with the other. This will ensure that your modem gets the maximum amount of signal. Don’t use radioshack/cheap crap cable amplifiers – they tend to cause more trouble than they’re worth.

Cabling and the quality of your coaxial cable splitters DOES matter A LOT! If you’ve got a splitter that’s gold plated, it’s probably crap. Check the back, if it’s glued on, it’s garbage. Quality cable splitters will have a back that’s been TIG welded on – like the ones that the cable company provides for you when they do the install. You can also acquire these from just about any reputable electronics dealer: monoprice, dalco, and parts-express being my favorites.

For cabling, you want to use RG-6 coax – not RG-59. RG-59 is not as well shielded and is basically inferior for data use – it was designed for use with cable TV systems when cable TV was all analog. RG-6 is a higher bandwidth cable with better shielding and should always be used with your cable modem, never RG-59. If you’re having stability problems with your modem, such as frequent disconnects or latency spikes, it could be your cabling or connections. Good RG-6 is still dependent on the quality of the connectors – using the screw-on gold-plated connectors will work, but it ultimately inferior to proper cabling done with a crimper. If you’re not comfortable with making your own cables or lack the tools (which you probably do if you got laid in college) then you’re best off purchasing the cables from one of the above mentioned dealers. Radio shack, best buy, circuit city, etc are often guilty of some heinous price gouging on cabling.

So there you have it, 4 ways to improve your cable internet access: Signal, Splitters, Cabling & Connectors. Questions? please leave a comment – I’ll do my best to answer your query.

I give you The Polar Clock

But what’s this? You want to hear about the preexisting fetish? Perhaps you expect to see llamas stacked in outlandish geometric fashion, each one wielding an exquisitely decorated placard bearing similar data to the above clock? You’ll have to keep reading to find out.

Nexxo’s Metaversa Project has served as a source of inspiration and a source of comfort. He’s the only one whose project has been dormant as long as my Goliath. His LCD screen setup (seriously, click and look at the pics) inspired me to no end. When I got Goliath’s VFD, I wanted a similar design and feel. With the help of Henri at lcdinfo and his custom VFD driver board, a bit of electronic tinkering, and LCD Studio I was able to make it happen.

Of course none of it is labelled, but I know what it means. The outermost ring on the left is CPU use %, the middle ring is coolant flow rate (it’s pretty much constant) and the inner most ring on the left is CPU core temperature. The history graph in the middle is CPU usage% history. In the upper right is chipset temperature, lower right is the % of maximum speed for the radiator fans (watercooled PC) and lower middle is video card GPU and video card RAM temperature. The big # that comes up in the middle of the GPU temp graph is the frame rate, as reported by FRAPS. I think I was sitting at the load screen for Counter Strike: Source when I shot this video. There’s also a break at about 1:00 where only one ring shows up – that’s the indicator for master volume and it only shows up for about 10 seconds after the master volume is adjusted, and I happen to have buttons on my mouse for doing exactly that!

So there you have it, influences and the interwoven nature of my tastes for data visualization. Even if none of this makes sense, or you didn’t “get it”, you can at least appreciate the opportunity you had to download an amazing and innovative screen saver.

I call up time warner cable and after about 5 minutes of wading my way through various voice menus and having my phone number read back to me after I’d input it… The CSR answers the phone and I told her what was going on. The box is stuck on E-13. She’s never heard of that error. She says did I leave it off for a minute… I think not but of course I say yes… That couldn’t possibly be it (it was in the end.) The box came back up just fine and I’m watching PBS-HD (Nature) now as I write this.

So what the hell is OCAP? And what do Mystro and OCAP have to do with the fact that my cable box (and possibly yours) lock up? Well, here’s the gist of what the CSR and I discussed on the phone while the torturously long proccess of an Explorer 8300 HDC damned boot sequence trundles along like an overweight crippled oompaloompa. Remember that description, it’s important for later.

The answer? OCAP, or OpenCable Application Platform is an open standard on which software for DVR functionality is run, much the same way my desktop and laptop run windows as a general purpose operating system. OCAP is half way between the hardware and the functionality of the DVR (or STB, Set Top Box, CPE or whatever) and the actual functionality (receving, displaying and timeshifting content.) This is why OCAP is known as Middleware.

So what provides the actual DVR interface? The front end, Mystro in this case. It even has a Public Blog but it don’t appear to be very popular at all. Why’s that? Because the box may have some nice features, but in a lot of ways it falls flat on it’s face.

This overweight oompaloompa of a cable box is born of OCAP – and apparently boxes designed like this have been mandated by the government, so they can get something that is very good for conumsers into the field: Cablecard. According to the cable company CSR (and I’m not sure I believe her), OCAP was pushed out the door to meet the cablecard deployment guidelines. It was “not able to be fully tested and still has regular problems.” Might be on par with windows, maybe a little worse. It’s sure as hell not linux. I’d love to blame this on the fact that the OCAP and the software it runs are probably (judging by the logos on the load screen) written in java. And you all will come to know that I hate java with a passion. It’s also possible that Mystro as a front end just sucks completely, too and that OCAP may actually be solid. I am unsure of the OCAP on the 8300 HDC is 1.0 or 2.0, but either way the 8300 HD (non-OCAP) is more reliable yet it is being taken out of service because of the new cable card regulations – a trade off that sucks for the consumer in more ways than it’s good. I don’t need cable card, not until it’s possible to just slap it into my PC, and would love to have a more stable box. Also, think of the waste of these boxes being pulled from service. Yuck.

So why is this a problem? This new technology (something we didn’t have in my area in State College) is wonderful, no PC involved, fairly quiet, functional. What does it lack? A lot of things. Firstly, the menu systems and interactivity in general are amazingly slow and sluggish for something that is supposed to be next generation technology. There is basic functionality lacking, things my XP MCE box has. Forward 30 second skip (probably abandoned because it makes it infinitely easy to skip commercials.) The fast-forward and rewind functions when resuming normal play have about 3 or 4 second response display on HD content. My PC (E6750, 8800GTX) can do this almost instantly, but i can’t very well get a cable card for it, can I? Thanks to the greed of the entertainment industry, we have DRM (Digital Rights Management.) You also can’t pull content off this box and archive it in a neutral format of your choice outside the DVR – something you can do with many other DVRs (which may require some hacking in some cases.) Also, it amazes me that cable co boxes haven’t caught up to where DirecTV boxes were 7 or 8 years ago. Once upon a time, I had a DRD-420RE and the menus were wonderfully zippy (although it was not capable of HD content, but neither was my TV at the time.) I’d expect the same thing from my box today, but I don’t get it.

The take home message should any cable company executives be reading this: quit kneecapping your equipment, enable and allow video to be pulled out the firewire port, develop better middleware or hardware or front end software any number of places the fix for the laggy menus and terrible unintuitive user interface (what the Fudge, A B and C buttons? – go look at a tivo or XP MCE remote!)

This has been a public service announcement, rant session, and the first post on my new blog. I hope for there to be many others like it.