If we follow the trend of other products, I would expect to see this in the 9.4 release notes:
* Removed "DISTINCT" and "GROUP BY". Usability studies show that most queries do not use them and new users find them confusing.
* "SELECT *" queries now return additional entries from Amazon.
* SQL language extensions to integrate Facebook and Twitter.
* Column order, if not specified in "ORDER BY", is heuristically determined from previous queries.
It's just great to see a release of anything that is actually better than the what it is replacing.
Dude, a little hostile there. You're a prime example of why this article even exists.
1. Yep, Customs. Not US customs. I worked in a top secret area without a security clearance as a foreigner, so security was actually frightening. Believe me or not. Basically, there was no security *at all* on really important stuff, like terrorist watchlists. A USB drive in a locked datacenter was the last thing in the world to worry about in this case.
2. I don't care if you store the thing over a sink with a 30 foot extension cord. You can kick it like a hacky sack for all I care. If we replace it weekly, that's fine.
3. Everyone had root access, as I mentioned before. Permissions were scrapped long ago, so that we could all manage build our little ship in our little 40 GB bottle.
4. Don't care about performance or reliability. Two developers couldn't even check out all the source at the same time, so if you had a day off, you're likely to find somebody deleted your home directory files just to have enough space to get a patch built. Zero reliability or performance already.
As I stated before, had you actually taken the time to read (and I don't get that actually reading, listening, or contemplation are strong points with you) it's not like I actually would have installed anything without permission. But when IT says that 8 or so people have to share a single 40 GB drive and that anything else is impossible, you start with the "I can get you 1 TB for $100" as a starting point for discussion. Sounds like you don't get that and aren't interested in doing anything but shouting down the people you think you have control over.
By the way, the head of the department did come around and asked my opinion. I stated what little I knew, and it got the ball rolling on an actual solution. The rest of the department apparently was too cowed be being called a moron or "shit monkey".
You explain the cost difference and why it exists. You explain why this may cause problems (stalls). You've thought through and explained the backup and concurrency issues. You present a second server as a viable option to this problem. You do not whine about how hard it would be or the paperwork involved. You are neither disrespectful or obstructive to the user.
Armed with this information, the Arrogant User (aka Desperate and Frustrated User) has a solution path and an explanation. He can now work with you and take it up the management chain to get something done. After all, he can explain his needs better than you.
I would want you as my IT person.
What part of having the entire qa, dev, and prod environments all sharing the same 40 GB not fill you with dread? 40 GB. Really, I'm not making that up, and this was a government agency. (Customs) They would rather have us doing "rm -rf" all over the place, as root, on the same filesystem that hosts prod. But that's preferable to you than just plugging in an external drive for scratch data?
Perhaps I just don't fully appreciate the complexity, but I have added an external drives and I didn't once have to repartition an array. 10 minutes was being generous. We didn't need a "set of drives", just something with more data storage than an iPod. I still haven't heard a reason why this would be so disastrous to even try - it does not fill me with dread. It never will. I guess I'm not as cool as you.
The fact that IT would just disdain a "shit monkey" and not even offer up a counter solution is *exactly* why IT departments are hated. It's not like I was going to break into the data center and start plugging in hardware. People like me only offer advice because we are desperate to get *anything* done. We don't really expect you to do what we say, and we expect with your uber-knowledge that you can offer up a better solution. Maybe prod you along and have you explain what you're really going to do to help and how long it would take. I can tell you would have provided nothing but excuses and obstruction, though, with that attitude. Good job!
The company I work for now, we all love IT. They fix things, they improve things. They say "no" at the right times, which is rare, and offer solutions instead of excuses. Night and day.
By the way, at Customs, the system fell over once while I was on contract there and the whole country briefly went to paper records for all ports and airports for all passengers and cargo, essentially shutting down the entire nation for the morning. But hey, at least there's no drive connected to the USB port, that might have caused real problems. Just keep using those 40 GB!
Here's my take on it, from somebody who really *did* have to share 40GB of space.
Arrogant User : "Our build server just keeps filling up. It's only got 40GB, you know."
IT: Would you please take a look and see if there's anything you can delete first? How about this directory that is for a 5 day old version of the product?
AU: No. We can't even do two builds at the same time, and all of qa, dev, and even prod shares the same space. We just need some space to even keep working. See, Fry's has them for $100.
IT: So Frys sells Ultra320 SCSI disks for $100?
AU: No, I never mentioned that.
IT: Yeah, except that your build server takes Ultra320 drives.
AU: So, just buy and external drive and plug it into a USB. I know the old server has them.
IT: We asked for the budget for a new server 2 years ago, and upper management denied the request, saying that spending thousands of dollars on hardware and a dozen or more man-hours migrating to the new hardware...wasn't justified.
AU: And they were probably right. You know what, I'll pony up the $100 and be back before lunch.
IT: That's nice. If we install it, it a)might not work properly since it hasn't been certified by the vendor and b)the vendor provides us with 4-hour turnaround, 24x7x365 support, but only for authorized parts bought from them. If your drive fails, they won't replace it, and we'll be blamed by management if we can't replace it fast enough and a failure occurs.
AU: Those are acceptable risks versus being stopped dead for hours every day. I'll buy a couple, then.
IT: Did we mention that if the drive fails in a year or two, it's unlikely we'll find a replacement? The vendor guarantees parts availability for these drives, or compatible parts, for several years.
AU: I said I'll buy two. Plus, I'm sure the $/GB will be even cheaper in a year or two.
IT: You also didn't think that if we can't find the exact replacement, we're rolling the dice, because different manufacturers have slightly different ideas of what "300GB" is. If other drives are smaller than your "300GB" drive by just one block, we can't use it to replace the drive, because it's in a mirror.
AU: So? We just need space. Forget about the mirroring stuff. Just plug in the damned USB.
IT: Great. Are you also going to pay for someone to come in during off-hours and do the swap, and then re-partition the drives? We're talking several hours of someone having to be in the office after-hours. That means overtime.
AU: I'll do it Sunday morning, right after the backups run. That'll be, what, 10 minutes?
IT: And you're going to justify the downtime to repartition on the build server to management, especially given that there's a release in a few weeks? If the drive swap-out goes badly, will you shoulder the blame for the delay which will strain relationships with our distributors and customers, and screw up profit projections by shifting sales more into the next quarter? And, will you shoulder the blame for 12 developers sitting twiddling their thumbs for 2 days while we rebuild the server?
AU: No, I'll just plug it in. No need to make this more complicated than it is. If it doesn't work, I unplug it and we forget about it. I'm sure your overwhelmingly-solid failsafes are more than adequate to handle the 0.00001% chance that this causes a problem. And by the way, we're pretty much sitting around idle anyway since we're all looking for files to delete and waiting for builds to finish.
IT: And you're going to fill out the change request forms?
AI: That's your job, right? I mean, that's *exactly* what you do get paid for.
IT: Yes, the change request forms your boss demanded we complete after we had an upgrade to your development environment server go badly, causing an unexpected 4 hour outage. Upper management agreed and we now have to document everything, have rollback plans, and get sign-offs from upper management and the manager of affected groups, which includes your manager.
AU: Again, that's exactly your job. It's astounding that you are resisting doing such a crucial task that falls directly in line with what you get paid to do. At what point is this bad enough that you actually *do* something?
IT: Screw you.
** AU goes of to look for files to delete and the entire team takes yet another 3-hour lunches while a build completes. **
Backed up the system lately?