May 2011 Archives

HP has been transitioning away from the cciss Linux kernel-driver for a while now, but there hasn't been much information about what it all means. Just on the name alone the module needed a rename (one possible acronym of cciss: Compaq Command Interface for SCSI-3 Support), and it is a driver that has been in the Linux ecosystem a really long time (at least in the 2.2 kernel era). A lot has changed in the kernel.

HP has finally released a PDF describing the whole cciss vs. hpsa thing.

Read it here:

The key differences:
  • HPSA is a SCSI driver, not a block-driver like CCISS
  • This means that the devices are moving from /dev/cciss to /dev
  • Device mode numbers will change
  • New controllers will increment kernel names, so a second controller will be /dev/sda, not /dev/sdb, so use udev names (partition ID, disk-ID, that kind of thing) to avoid pain.
  • For newer kernels (2.6.36+) cciss and hpsa can load at the same time if the system contains hardware that needs those drivers.

The news was hot in the office when Symantec announced they were eating buying Clearwell, another eDiscovery firm. This is big news since we compete directly with Clearwell. The Register has a really nice article about it, and give a nice summary of WTF we do.

Read it here!

It was also the same day that Gartner released its very first Magic Quadrant report on the eDiscovery field. Coincidence? I think not.

Logik wasn't on the chart since we didn't meet the minimums. But anyway, it's a good thing to read since decision makers put a lot of stock on Gartner reports, and gave me a good feel for the industry as a whole (which I had been lacking).


Anyway, perhaps the eDiscovery market is entering its consolidation phase. Don't know.

The power of nice

John Scalzi, SF author, once quipped:

The failure-mode of clever is asshole.
Indeed, he is not wrong. I bring this up because when sysadmins get together and a new person comes along with really basic, not researched at all questions, the response is often sarcasm and mocking. And sometimes riffing on the mocking in increasingly clever ways.

The ServerFault chat-room transcript has examples of this scale most days. Most often it's the "Jeez that guy is a putz" kind of gripe, but sometimes it expands into more then that. Compared to IRC, ServerFault is a very nice place, but we still have our moments of withering scorn. As a moderator I see more of the withering scorn as those answers and comments earn flags (yay!). However, even when folk are being nice it can be very discouraging to new entrants.

An example is this question.

The asker is looking to set up a server infrastructure for an undisclosed application and asks several questions around that topic. It's clear from the question that they haven't dealt with this kind of problem before.

The first comment states this kind of question can't be answered here and gives a very broad reason.

The second comment, by the author and after the question earned a pair of down-votes, asks for down-voters to explain their reasons (a common gripe among new-ish people who are earning downvotes, as it happens).

The third comment ups the snark ante and denigrates the experience of the asker in the process.

That's where I found it this morning. Rather than let it get closed as 'argumentative' or let the comment-snark get out of hand, I dropped an answer on it explaining, in detail, why this kind of question is not able to be answered on ServerFault and what more information would be needed to answer it at all. I used neutral tones. It did the job, since the snark stopped and as of just a minute or two ago the asker marked that answer as accepted.

Even though the full extent of their question couldn't be answered, by taking the time to answer what could be answered, and describing what is needed to answer the rest of it, the original poster learned things. And importantly learned more than just, "you don't know enough yet, come back when you do."

Sometimes, you just gotta expand "It depends."

In the comments, Holocryptic pointed me to an article I fully intended to link to but couldn't find. Kyle Brandt, of ServerFault/StackOverflow fame, posted to the ServerFault blog back in October 2010. The topic touches on a lot of what I mention above.

Go read it.

Tools of the trade

I made a few minor hardware purchases recently. I have a new job, and they've never had someone like me full time before, so they don't have the kind of tools that a system engineer just accumulates over time. I'm slowly fixing this as I run across things I don't have.

So far:

  • A laptop with a DB-9 serial port on it. Because even today networking gear needs RS-232 to be configured.
  • A laptop with enough RAM to run multiple VMs. Because staging virtual-machines is part and parcel to what I do, and Logik doesn't believe in Desktops.
  • A docking-station/port-replicator for the laptop. Because I don't want a forest of wires plugging into the laptop where I can see it, I want the forest of wires behind it. Also, plugging all the USB and other power-drawing stuff back there means less heat generated on the laptop itself, which increases its lifespan.
  • An ExpressCard GigE adapter. Because having two network ports is occasionally very handy, and the built in one on my laptop doesn't like promiscuous mode for some reason. I need me my promiscuous mode. This card also supports a larger Ethernet frame-size than the built in one (9000b vs 4000b), which is useful for iSCSI work.
  • An ExpressCard eSATA adapter. Because in our business sector having 2 such on a device is occasionally very useful. Never used eSATA at WWU, and here we're regular users.
I'm sure I'll add more. They already have things like cable crimpers and the fiddly tools needed to flip hard to reach dip-switches or get into things with very small screws. They are missing torx-head screw-drivers, but I'll order those when I run into them again. At least at this job I'm no longer having to occasionally dismantle SDLT drives to extract tapes.

A rocky transition

It has been 6ish weeks since I started at Logik and I'm finally getting my arms around the special problems this particular workplace offers.

It has been hard, no bones about that. Moving from an organization with 4000+ employees to one with well under 100 means that I'm moving from a pretty well siloed IT department to a one-gun setup. I'm doing things I've never had to do before, or haven't had to do for 7+ years and am therefore rusty.

My horizons, see them broaden.

Even though this is exactly the kind of thing I needed, it's still making me grumpy.

Grumpy? Me? Yes, it happens. It happens more often when I feel like I'm being held to unreasonable expectations. And at my old job, what I'm doing right now would be unreasonable expectations. But, you know what? I did this on purpose, so I should just buck the !%$!^$ up. I have to keep reminding myself of that lest I snarl at people.

Like most geeks I know, I get, ahem, testy when my ignorance is underscored. We have pride in Knowing Things You Don't, and finding areas where there are vast fields of unknown things.... well, defensive strategies start getting deployed. The well socialized among us hit the books, figure it out, and come out the other side educated and competent with a minimum of fuss. That's what I'm trying to do.

So what kinds of "unreasonable expectations" am I running into?

OS X support.
We have Mac users, quite a few of 'em, actually. Part of our recruitment package includes a choice of PC or Mac for doing your daily work. So long as you can get your job done, who cares what you run? These users rightly come to me when things go wrong and their own google-fu has failed, and so far I've been mostly useless at this. Unfortunately, mac users are used to this from professional IT. I still hate being mostly useless.

Laptop rebuilds.
I haven't done straight up desktop support since 2003, and even then it was very occasional, so this kind of thing simply takes getting used to. I can do it, no worries there, but still. Part of me is convinced it is a step backwards, but then I club it with the "on purpose" line I mentioned earlier and it sulks back into its cave to mutter darkly.

Network hardware configuration
The last time I was routinely in the cli of switches was in 2003, and even then all I was doing was setting up span ports for sniffing. I set up a ProCurve switch for iSCSI at WWU late last year, but that was a pretty simple config. Actual VLANs and stuff? Never. I know how it all works in theory, but converting theory into command sequences is another thing entirely. The fact that our wireless controller is no longer broadcasting anything even though the interface says it is still has me stumped, and I hate that. Happily, we had exactly one wireless user and we got a cable to her, but it's still annoying not to have in-office mobility.

I'll get there. I went through a similar grumpy-WTF period when I started at WWU, so this is all part of the process.