I migrated yet another GroupWise post office from NetWare to SuSE Linux this evening. However, I had undercover help.

The data on the NetWare volume was 65 GB. So we made the ext3 volume 100 GB. Did the initial copy, and had 1% free disk space left. Dang NetWare compressed volumes....

Time to expand the LUN for the new server. The destination box is a Xen VM guest. I'm shady on the details, but after the LUN was altered on the host VM, the entire blade has to be rebooted to see the new space. The admin who was going to take care of it, was going to do it early in this morning, as he normally shows up at 6:00 AM. About midnight, he wakes up: "Oh heck! I'm scheduled to come in late tomorrow!"

Still in bed, he grabs his BlackBerry and establishes a SSH connection to the host VM. Resizes the LUN from the command line and then reboots everything. All is well for me to get the job done tonight (which I did, and it went well).

I did say "undercover" did I not? ;-)

Here is an analysis of where the takeover bid by Elliott Associates, L.P. is going. It is long, but seems to me pretty well thought out. It points our that this is a loss for both Microsoft and Red Hat. And a huge loss for Novell (employees and customers).

In my last JE, I asked how to back up a directory structure with way too many subdirectories and files in it. One of the ideas was to do one tar job per top-level directory. Since this one didn't take any more disk space, the SAN guys wanted me to try it first. I had to learn how to use Perl's Parallel::Forkmanager. It worked well.

I have 256 top level directories, so I forked one tar job per directory. We took the VM that runs the server and configured it for four CPUs. I configured Parallel::Forkmanager to only allow four concurrent forked tar jobs. What was a 70+ hour single tar job became a 30 hour batch of 256+ tar jobs. The new scheme runs in 43% of the time of the old job. That's great.

Thank you for the advice. I'm using it well. :-)

I have a system that makes a metric buttload of small files, with a 1:1.2 ratio of files to folders. So 850 GB of files in one million subdirectories. The tape backup system chokes on this, and I have the disk space to copy everything into one big file - the tape system loves backing up that. As my storage grew, my initial .tar.gz job got to taking three days to complete. I thought that converting it to just a .tar file would speed it up, but the opposite happened. That may be due to the fact that I changed the destination file system at the same time.

Any suggestions for an optimal setup to copy a large number of small files into a single file for tape backup? The tape backup exists for disaster recovery, in case the SAN burps (again). I have a lot of flexibility on how to get the job done - suggestions welcome. If there are any system tuning parameters I ought to be looking at, please say so.

Thanks!

I had installed The GIMP on my wife's machine way back when. Version 2.2 on Windows. She wanted to scan and crop photos, and it seemed a good way to introduce the idea of free software to her. Free, as in hobbyists are writing it and giving their work away for free. The GIMP can do an awful lot, and my wife could see that we didn't have to pay a lot of money for an app that duplicates that The GIMP does. And, you can see that the world is just a little bit nicer, that people do good works and share it.

Alas, I had to work on her machine a week or two ago. While I was there, I upgraded to The GIMP 2.6.

The GIMP 2.6 does not do "Acquire from TWAIN'. Heck, that was the whole reason she used The GIMP.

Nicely enough, the downgrade was blindingly simple. But she will have to remember to never ever upgrade.

A few weeks ago I (and it seemed a number of other people) were dismayed that the link on the front page to the message center disappeared. I found it's been put back, though in a slightly different location. Now, underneath the list of new messages is a link "NN more" which takes one to the message center. It's on the right instead of the top center.

I know that I complain a lot about the way Slashdot is run. In this case, I want to specifically say: THANK YOU. Thank you very much. Taco and friends do deserve thanks for listening to their content-creation-base and acting on it.

For a while now, I've been looking at the behavior of organizations to see if I can identify where it all went wrong. The vendor I have the most to do with is a huge glaring example. Slashdot itself repeatedly gets it wrong. The thievery at hospitals, ripoffs in the mortgage industry, the slow self destruction of the USA automobile industry - it all seems to point to the abandonment of the company motto:

"It's good to be a (insert name) customer".

If that were the standard by which organizational decisions were made, we wouldn't have all this trouble.

The Novell company motto seems to be "It sucks to be a Novell customer (and we're OK with that)". They do some very good technology, but insist on overpricing it (because it is NOT GOOD to be a Novell customer) and then wonder how they are going to pay the support bill for a product with 1% market penetration.

The Slashdot company motto seems to be "We'll write the code the way we want to, and you all can piss off". Yes, as a matter of fact, I liked the front page to have the one line messages status text and link to all my pending messages. Posting this JE took far more work to get to the 'create a journal' link than in the past. But I don't expect it to change, because "It sucks to be a Slashdot content provider (and we're OK with that)".

Golden West started it's wholesale ripoff of people by handing out first ARM mortgages, then NINJA mortgages. By that time, their company motto had become "It's stupid to be a Golden West customer (but we're OK with that)". It's worth noting that the banking crises did NOT affect most of the small town local banks who don't view their customers as prey.

When was the last time someone bought a new Chevrolet and said to themselves "Well, at least I didn't get stuck with a Toyota!" (because "It's sad to be a bailout car customer (but we're OK with that)").

There are countless organizations where the company motto is "It's good to be a (insert name) shareholder (but we play our customers for fools, and we're OK with that)".

It's few and far between, but:

"It's good to be a Costco customer".
"It's good to be an Amazon.com customer".
"It's good to be a Netflix customer".
"It's good to be an Apple customer".

Why is it so hard for organizations to fight the decisions that transform themselves into a "It sucks to be a (our name) customer"?

Yesterday I noticed that my GroupWise mailbox was snappier. Didn't know why, and thought that maybe the email archiving software had trimmed my mailbox (which would be a good thing). I've got 19 GB in my mailbox, and 18 GB of it ought to relegated to the archive. Nope - turns out my co-worker had added RAM to the server while nobody was watching.

GroupWise has a mode called Client/Server where your desktop fat client talks to the server via TCP/IP packets. Back around GroupWise 5, if the post office agent (on the server) went down at all, everyone knew it because every machine immediately got an error on the screen: lost contact with server! Somewhere around GroupWise 6 or 6.5, Novell reprogrammed the fat client to wait instead of immediately throwing up an error. The timeout is about 30 seconds.

Last year we started a whole push to virtualize as much as we can. We use Xen virtualization, as that is what Novell supports and includes in SuSE. Paravirtualization has a few advantages, one of which is that you are essentially running a guest OS that is the same as the host OS. So when it comes time to reboot, the OS is already loaded in RAM - the Xen host just has to create the guest, and do a bit of linking. We typically see an init 6 take ten seconds from time we lose continuous ping packets to the time we get them back.

My co-worker was looking at the server stats, and the mail server showed pretty much 100% utilization 24 x 7. Certainly the post office seemed slow. He thought it might be a process gone bad, but no - the mail server was just that busy. It looked like it could use a little more RAM though.

(Prepare by opening the Virtual Machine Manager)

Init 0
Edit the VM definition to have another 1 GB RAM
Init 3

It is up, and before the 30 second timeout from 120 fat clients. Nobody notices, nobody calls. It just works (and faster, too).

Sweet. :-)

I'm an old guy (my first program was in FORTRAN on punched cards in 1979) and I learned programming the old-fashioned way. Which means that I like my source code to be obvious as hell.

Here's what makes sense to me*:

$source = "abcdef";
$pattern = '/^def/';
preg_match($pattern, $source, $destination);
print($destination);

Really, what I want is:

$source = "abcdef";
$pattern = "/^def/";
$destination = regex_match($source, $pattern);
print($destination);

How it's done in Perl:

$source = "abcdef";
$pattern = '/^def/';
$source =~ /($pattern)/;
$destination = $1;
print($destination);

Why do I hate thee? Instead of letting me assign something to $destination, I have to encapsulate my regex pattern within parenthesis (because OF COURSE that's the assignment operator), and now the result will magically be assigned to variable $1 - wherever the hell $1 came from. THEN I'm allowed to assign $1 to $destination. Wonderful. And by the way, the single quotes around the regex pattern matter - you cannot put it between double quotes.

Obvious as hell, it ain't.

*Note that first snippet is working PHP code. Yes, I would have answered my own rant, except that my postfix and GroupWise boxes do have Perl on them, and do not have PHP on them.

So no-one wants to maintain KRegExpEditor which was a part of KDE 3 utils. So it was dropped out of KDE 4. Turns out the least hassle way for me to run a local regex tester is to:

1) install Wine
2) install Microsoft Visual C++ 2005 Redistributable Package (x86)
3) install The Regex Coach from http://weitz.de/regex-coach/#install

gnome doesn't seem to have a regex buddy either. Yes, there are a ton of web sites that will do the work - but I don't really want to work like that.

Well, it is better than cranking up the xen virtual machine I have of Windows XP Home.... Cut and paste is far easier.

Today was GroupWise upgrade day. We upgraded 30 servers from version 7.foos.bars to 8.0.0 (+)*

It was almost a perfect. By 14:00 everything was done except the one WebAccess server. Turns out I pointed at the wrong (older) version of Tomcat during the re-install** and it took a while to figure it out. I had to ask a co-worker to find the problem, because I don't yet grok where all the Apache and Tomcat pieces live, especially since they move locations when you advance your OS to newer versions. This was an older OS with the new Tomcat, and I missed it.

The new WebAccess has AJAX-y goodness now, so when you scroll the mouse wheel, the web server fetches more data to display without a page reload. It's subtle, but much nicer. The web pages act a whole lot more like the fat client now.

The real star of the day is the Linux fat client. It *finally* looks and acts like the GW 7 Windows fat client. It's got the customizable home view, color codes based on category, implements custom tab views. I didn't see anything missing from the GW 7 Windows client, although I think that there was a least one feature that didn't make it. I think it was the DMS, which we don't use anyway.

It now has a Notify feature that can pop up when you get new mail. That's been in GW since version 4 (or before) so it's good that it finally made it.

The GW 8 Windows fat client did get some new features, but I haven't played with them yet. I was just happy my settings and custom applets worked as if nothing had changed. The best kinds of upgrades are the painless ones. ;-)

*I've got pre-release service pack code, so that I can test a bug we found a little while ago.

**This is the one component I wanted to rip and re-install, so that weren't any legacy pieces of code still getting loaded. If the thing were to ever flake out, I'd have that remorse of not doing it right the first time. So I spent a little more time today in hopes of less maintence problems tomorrow.

I have no idea if this link is persistant or not; but, at the moment it shows that Digg has surpassed Slashdot in the Netcraft toolbar ratings. Slashdot is the #513 most visited web site on the planet, and Digg is #512. Google is still #1.

At one time, Slashdot was ranked #25 world-wide by Netcraft.

It will have fallen to The Onion less than 30 ticks from now.

Of course, sites like Facebook, MySpace, Bebo, Friendster, YouTube, Engadget and Blogger bumped the 'dot out of the rankings by going after larger audiences. Although I don't know of anything that made it impossible for Taco and friends to cater to those people.

Other sites, like Weather.com, Foxsports, Wikipedia, Drudge Report, Woot, - they all blew past Slashdot - but they aren't really interactive content sites. They publish, you read. It's a different paradigm. The 'dot is all about community.

The fundamental question that is left: how long before Slashdot is less popular than I Can Has Cheezburger?

100 ticks and counting....

You Digg it?

It's been a long time, but GroupWise 8 has finally shipped. The article gets one thing wrong - GW 7 had a Home View; what's new in GW 8 is that the home view now shows up on the other clients as well. Speaking of which, what's really new with GW 8 are the additions to the Linux client. In the beginning of the beta test, Novell was told that, as a Linux company, their groupware product should treat Linux as a first-class platform, not second class. Yes, GroupWise runs on the back-end on Linux great (and has for years) - but the client was less than impressive. If the GW 7 Linux client was 80% feature complete (compared to the Win32 client), then the GW 8 Linux client is 95% feature complete. WebAccess got a make-over too. Don't know about the Mac client though. Enabling busy search of published calendars via URL should be interesting.

Oh, and a slightly funny spin on a google blog: Radicati Study Sponsored by Google Shows Microsoft Exchange has 57% More Downtime than Novell GroupWise ;-)

Today (minus twenty years) WordPerfect Office 2.0 for DOS LANs shipped. WordPerfect Office got renamed to GroupWise in 1994; here we are today.

Although we got chided a little bit by some Novell staffers. Today marks the 20 year birthday of the GroupWise that runs on DOS - but the original product was released for the Data General mini in 1986. And it ran on a VAX/VMS before it did DOS. Isn't celebrating the DOS version is a rather Microsoft-ish thing to do? <hangs head>

Anyway, back in 1998, someone wrote A History of GroupWise. It's terribly sales-pitchy, but might be a fun read. Did you know that GroupWise was used as the messaging system at the 1994 NetWorld+Interop show in Las Vegas, Nev., where approximately 88,000 users sent/received 2.2 million messages? Or that WordPerfect Office 3.0 was the first LAN-based, multi-server product to offer E-mail, calendaring and scheduling across multiple platforms?

The product manager for GroupWise, Dean Lythgoe, posts in his blog that he still has three people on staff from back in those days.

Anyway, I really like GroupWise. I thought it was so cool, back in 1996, when GroupWise would send a message to my text pager fifteen minutes before every meeting (I had set up the GroupWise Pager gateway that used a modem to dial AT&T). Once or twice, I'd been sent out in the field, and this feature saved my butt as I high-tailed it back to the office, just in time!

Anyway, Happy Birthday (to the DOS version) of GroupWise. :D

Today, my quantity of GroupWise servers went from ten to nine. And with that, my GroupWise system is running 100% on SuSE Linux. We powered off our last NetWare GroupWise server.

It took a while to get to this point. I've been 90% on Linux for two years now. But our old anti-virus gateway ran on NetWare only, and moving people to the new anti-virus (plus anti-spam!) gateway was never highest on the priority list. Last Friday, we flipped the switch, and everyone goes through the new system.

You may recall that I'm in an organization that is merging with another. Likely at some point, I'll be the GroupWise admin for their boxen too. They have 20 GW servers on NetWare. That conversion will be a nice big fat project to look forward to (assuming I don't get moved to something else).