Bloggy Stuff
So, Duncan is celebrating his tenth blogiversary™ at Eschaton by counting down to the worst wanker of the decade.
John Mckay celebrated his ninth at archy with just a passing mention, as did Jams O’Donnell at Poor Mouth for his sixth.
Moi at Bloggg decided to shut up shop and move on to other things.
We are told that blogs are over, but I doubt that, mainly because everything that is supposed to replace them keeps coming and going. The problem is the urge to change things and make them ‘better, faster, cheaper’, which is great initially. Unfortunately, people don’t know when they have hit the optimal point for utility and tend to change for the sake of change.
Jerry Pournelle who has been writing about computers since the days of CP/M has a useful standard for things – Good Enough. This means that something makes things better. It isn’t perfect, but it is useful and worth having. Too many things that were Good Enough have become a PITA through ‘improvements’. [Did you know that Blogger is going to version 3? I’m at my current location because of what they did to me in the change to version 2.]
Oh, well, back to sniffling and sneezing…
18 comments
“Did you know that Blogger is going to version 3?”
I didn’t know you were originally a Blogger blogger. WordPress, even the free hosted version, is in general much less hassle than the current BlogSpot-hosted Blogger, but even WordPress.com has its problems. Last week, for instance, my Yellow Doggerel Something lost its yellow for most of a day when a developer made one half of a two-part change to the way backgrounds are applied. That result indeed wasn’t Good Enough, and I actually filed a trouble ticket (I even suggested where to apply the fix). To their credit, they fixed it a mere few hours later.
The reason I switched to WordPress was the ability to import my BlogSpot archives. I lost my Holoscan comments, but everything else came across. It looks like Echo is about to die, so even those who stayed with them are about to fall.
About the only vestige of my original setup is the SiteMeter counter.
I don’t get too involved with the WordPress.com annex that I set up unless something happens to NFS and I’m in the middle of one of my passions [hurricanes, or one of the sporting events I follow], but it does let me mess around with new versions before I update this site.
When they went to Blogger 2.0 they locked me out. I have no idea what happened, but I couldn’t get to the site as an administrator, so I shifted to this site earlier than I had planned.
They all have a tendency to change things at very inconvenient times, so I’m much happier out on my own.
“They all have a tendency to change things at very inconvenient times, …”
They all have a tendency to change things that work to things that don’t work, and deploy the result with inadequate testing.
I have no great stock in my old stuff; I leave all of it up in its original form (pile-of-pages; Blogger; WordPress) rather than trying for continuity. All my HaloScan comments vanished, but I did download them before they did so, so if I really have a hankering to find a particular old comment (with no context), it’s there in a file in some XML format. When my life is at an end, I’ll wish I had back all the time I spent changing blog platforms.
No way was I going to reward JK-Shit (or whatever it was) by moving to Echo. In my vocabulary, Echo was a large silvered reflective Mylar satellite from my childhood, visible occasionally from our front yard (Space… the final front yard!)… but not something for 21st century communications.
I have been very slow to upgrade to the newest version of WordPress because it required me to update the versions of MySQL and PhP that I use on the server. I did it a while ago and have been waiting for something to break. I was also turned off by the number of upgrades that occurred to the version I was initially planning to upgrade to, so I let that revision lie, and will skip it for the more stable new version. I never upgrade to rev 0.
Almost every upgrade I’ve seen are designed for features I don’t use, or want to use. I’m still happy with text and few pictures.
I was forced to upgrade pretty much everything at the office — web server, database server, NIS server, DHCP server, name server, bug tracking system, source control system, everything — because our antique equipment started dying and the antique versions of Linux and associated software wouldn’t run on modern equipment. Luckily it’s a pretty well known process for doing all this — mysqldump, mysql import (after turning up the max command line length, important!), etc. — just a PITA. Especially a PITA was the iSCSI initiator issue, Red Hat switched from the (slower than slow) Cisco initiator to the (much faster) open-iscsi initiator between RHEL4 and RHEL5 and the open-iscsi initiator is a royal pain to manage, they took user interface design hints from mdadm (GAH!) — which is sorta a problem if iSCSI is your back end storage. But anyhow, after totally re-implementing everything basically from scratch (just importing the data from the old junk hardware) and turning three racks full of antiques into 8U of servers and storage that provides far more capability than those three racks full of antiques provided, I remember once more why I quit running a blog on my own software platform and decided to just let Blogger do the blogging. Blogger is a PITA, but less of a PITA than the alternative. Sorta like Linux sucks, but sucks less than the alternative. I do this all the time for work, I don’t want to do it all the time for play, sigh!
Oh, yes, I remember not so fondly being a systems admin when everything started to die and you have to change everything because no one would approve a reasonable upgrade program, but always found the money when the entire structure started to collapse. Of course that was a problem for the entire university system, where major portions are funded by grants, instead of being part of the regular budget process.
I don’t have those responsibilities any more, so I have time to screw around with the blog. It is just a short trip down memory lane, not the entire nightmare it can be when you have people interrupting to ask if the system is going to up at some point before the time stated in the memo they received about the need to do this work and the time table to complete it.
What is amazing is that I took two racks of storage arrays, and put their contents onto one 2U storage array, and I still have 10TB of free storage left. I remember when a terabyte of storage was a Big Deal. Nowadays you get two or three of’em on a single drive. Or you would if Thailand’s factories hadn’t washed away, it took us literally days to scavenge up enough 2TB drives from practically every distributor and vendor on the entire West Coast to fill that 2U chassis, forget about 3TB drives. Globalization. Yay.
Oh yah, the old array would have cost somewhere around $500,000, and had an awe-inspiring six terabytes of storage. The new one has twelve 2TB drives (24 terabytes) and cost around $6,000, we can add up to four more twelve-drive JBOD’s to the thing for a total of 120 terabytes (and you don’t have to shut down the storage box to add a JBOD, just plug it in, the storage software notices the new drives, fire up the storage manager, tell it to use the new drives as a twelve-drive RAID6, voila it’s now added to the storage pool). So I don’t think we’re going to run out of storage anytime soon (though this is RAID6 so basically two parity drives, but still). Oh yah, the reason I have two of’em is for mirroring — sure, each of them is RAID6, but if disaster strikes I can theoretically flip over to the mirror within thirty minutes or so and have the company back up and going. Not that I expect that to ever happen, but… (shrug). As cheap as they were, why not? Just turn on the replication on box A to tell it to replicate everything in real time to box B, voila.
– Badtux the IT Penguin
The initial hard drive unit in the first minicomputer I administered was $10K, and it had 5MB fixed and 5MB removable using 14-inch disks. Backup was a 9-track tape unit. My 5MB, 10MB, 20MB, 40MB, and 80MB Winchesters on my micro each cost $900 when I bought them, and that was the wholesale price.
Yeah, if you are operating a business these days, there is no real reason not to have redundancy, because even at minimum wage recreating that data would be obscenely more expensive than the cost of a disk array.
It sounds like you’ll make a 5 year pay off on utility savings alone, given the comparative electrical and A/C requirements of the two systems.
I wish. Our monthly electric bill is in the six figures, and our office is *not* that big. But we have two large labs and two small labs each of which has rack upon rack of servers and / or security cameras (I haven’t counted the number of security cameras we have, but they’re surely in the hundreds). We have dozens of custom test jigs where security cameras get to watch Mandelbrots for load testing purposes, as well as security cameras scattered around all over the place just to get real-time data to analyze, as well as a rack full of camera simulation equipment to simulate hundreds more cameras…
Which reminds me that some moron decided to break into our office and steal some stuff. We have his face. We have the make and model of his car. We have his friggin’ *license plate*. What kind of moron breaks into a VIDEO SURVEILLANCE company?! ROFL!
— Badtux the Easily Amused Penguin
My first guess on the perp is that he had no idea what ‘surveillance’ means, and thought you had DVDs. There is a low level of literacy among the criminal class. I always felt that passing the GED should be the first requirement for parole or probation.
Actually, your marketing department should be all over it for a presentation to clients. Hopefully the local PD will actually do something with the information you gave them, and not just sit on it.
I was just thinking of the cost of operating the stuff you replaced, and its effect on electrical usage. Video is really a bigger energy user than most people realize. It is sort of shocking how much power flat-screens suck up, and CCDs are as bad as LEDs for energy consumption.
Well, he stole some CPU’s and memory and other high-value small stuff that was easy to stuff into his pockets, so clearly he wasn’t a *total* moron. Just mostly one :). As for whether the Sheriff’s Department got the dude, that’s above my pay grade, I helped get all the relevant stuff off the video management system onto a thumb drive to give to the deputy after he viewed the relevant sections of video with us, any interaction with the Sheriff’s Department after that would be through the business manager.
There are four new 1U servers, each of which are dual processor with at least 12gb of ram, and two new 2U storage arrays, each of which has twelve drives. The 1U servers have dual redundant 350 watt power supplies. The 2U storage arrays have dual redundant 400 watt power supplies (they have twelve drives, but only one processor, and Xeon 5550’s use up to 95 watts of power so leaving one of the processors out saves a lot of power budget — note that all of these machines have identical motherboards, keeps our repair inventory costs down to do it that way). So total draw will be 1400+1200=2600 watts. Vs. two racks worth of equipment, or a total power budget at least 5 times more. Uhm, yeah, we’re saving at least 10,000 watts of power, which at PG&E rates is pretty darn good (that’s $1.70/hour that we’re saving, let’s see, 24/7 for 365 days a year… $14,800/year savings, dayum, two years and it’s *more* than paid for!).
Man, I need to go measure these numbers exactly and run’em past my boss. Maybe I can get a pay raise or somethin’ out of it ;).
I was certain there had to be a savings because I have seen the same thing at other places when they upgraded. Those are some significant savings from one segment of the business, and it is also a marketing item for upgrading customer systems.
Getting hard numbers is always the best way of dealing with it, and it helps keep any sales promises within a realistic range.
Heh. Ran the actual numbers yesterday and sent’em to my boss. He sent them up the chain, and kudos from the CEO filtered back down the chain to me. That evening I did the final switchover from the legacy storage arrays (basically a full rack of equipment) to the new storage arrays (two 2U units), they will be decommissioned next week when I’m physically in the office (VPN, ESXi, IPMI sorta makes it unnecessary to be physically present nowadays unless a cable needs to be moved or unplugged). Everything worked smoothly and without a glitch. Friday the 13th was a very lucky day for me ;).
It is always good to have your name known favorably in the upper levels, especially when it is attached to cutting costs, which appeals to the financial types that seem to gather at the top of the corporate structure.
All too often IT is considered overhead, even in tech firms. It is almost as if people don’t understand the reason they make money. The top of too many corporations feel that ‘playing with numbers’ is the most important thing a corporation does, rather than creating what customers want to buy.
I put it down to hypoxia in the upper management because of their office locations.
Our CEO is a salesman, not a finance man, so he’s not quite so myopic as some of the folks at the top. He’s more of a Steve Jobs type than your classic beancounter, though his reality distortion field isn’t anywhere as good as Steve’s was (but whose is?). That said, he has to sell the VC’s on continued funding for the company or manage to turn a profit so he doesn’t need VC’s anymore, so saving a sizable chunk of change is decidedly appreciated. And yes, I was *quite* surprised to find that this had filtered all the way up to the top with my name still attached to it, you know how that credit thing *usually* works in companies, I’m sure :).
Oh, you mean the bit about the press releases for the promotion of your boss’s boss that somehow include a list of accomplishments that came from you or your team, yeah, I’m familiar with that type of ‘misunderstanding’ in large companies.
Sales, so the CEO still knows what it is that the company makes. That’s always helpful, and is increasingly rare among large corporations where the CEOs tend to think the corporation just sells stock.
Sales, so the CEO still knows what it is that the company makes.
Boy does he, he’s focused like a laser on that, and he lets us know what the people he talks to *want* us to make too, some of which is possible with existing technology, some of which would require Star Trek physics in order for it to work :). But anyhow, yeah, I’m quite happy with where I’m working right now, we have the usual problems such as various legacy issues to work through but we don’t have the problems that are misery-inducing, if you know what I mean.
It is always a much better environment when management knows what is actually going on. The conversions are much more realistic, and they tend to understand when problems are explained. The misery is when you tell them what can and can’t be done, and they refuse to accept that some things just can’t be done, no matter how much they may want them. All of them quote unrealistic time frames.
I’ve seen too many RFPs that were pure imagination, no grasp of reality, and then had calls asking why I didn’t respond. In some ways the worst are those where you can do what is asked for, but you know it isn’t what they need, and they will just be PO’ed no matter how good a job you do.
And people wonder why I wouldn’t work for government agencies or attorneys. The Fortune 50 corporations I did occasionally work for, were bad enough, and very slow paying, I might add.