I’m currently running 10.0.3.2 of ZCM and am prepping to make the move to 10.1.1, which will require a DVD install. I want to make sure that my stuff is backed up before I do this! Here’s how:

From the ZEN server (on Windows)

1) zman admin-store-credential administrator

2) zman database-backup c:\dbbackup (or other path)

More on scheduling this later.

There has been a lot of buzz lately about the long-awaited release of SP1 for Vista. There has also been no shortage of press articles about how Microsoft has released a dud of a product, and how everyone should wait until SP1 comes out to get serious about using it, especially businesses, where the potential for a headache is greatest. Stories abound of users switching back to XP after fooling with Vista, and being oh so much happier!

The FUD abounds.

I’ve been running Vista for a little over a year now, and have been quite happy with it. I’ve even had several computers in that span of time, with varying hardware configurations, and they all took Vista quite well. All of my Apps and Games run about the same as they did on XP, some a little bit faster, and some a LOT faster.

But just for grins and giggles, I decide to go ahead and take the plunge and install SP1 for Vista, even though it has only been out a short while. If I toast my machine, so be it. Maybe I’d be able to learn something interesting from the whole affair, and I would at least be able to post it here to help out somebody else who ended up with the same problem.

So I proceed. I just let it download the update through the built-in Update functionality of Vista, as opposed to downloading the redistributable package. After I downloaded and installed it, I rebooted it, only to find…..

Everything still worked. My apps still worked. My games still worked. My printers still worked. File copies were a bit faster. Boot time was at least the same as it was before. I didn’t have to download any driver updates for my relatively recent hardware. In fact, the whole Aero interface seemed a little snappier, and my usual apps loaded and closed just a tad quicker. No problems to contend with whatsoever. Oh and BTW, I’m running Vista Home Premium 64-bit…. not that this matters much today, but not so long ago, if you ran Windows XP 64-bit or just about any 64-bit version of an OS for that matter, getting drivers and driver updates was somewhat of a crapshoot. Looks like Microsoft has done something right.

I feel like this is a typical experience for most Vista users updating to Service Pack 1. I’m sure there are some users out there whose machines were torched by SP1; but this sort of thing has been typical in the IT industry from it’s inception. Sometimes when you install an update / service pack for a product, it WILL blow up in your face. It’s a fact of life as a technician/engineer. If Vista is your first experience with this, then you probably have not been doing this for long. The only difference now is that Microsoft and Vista are under the microscope, and it’s producing all sorts of negative publicity, which feeds the FUDmonster.

I’m getting better, actually. I haven’t posted anything on here yet for 2008, but I promise that I’ve got some stuff coming up. Stay tuned.

Pretty simple, but slightly different:

write net running-config 1.2.3.4:config.txt

where 1.2.3.4 is the IP of the TFTP server.

Well I just got back from a week-long vacation in the carribbean last week, and I’m finding it mighty hard to get back in gear.

I haven’t posted much on here lately, especially not anything particularly interesting, just a couple of opinions on this and that. The truth is, I haven’t really done much lately worth reporting on; no cool new little utilities to brag on, no scripting tricks or around-the-elbow fixes. I’ve been reading TheBackRoomTech’s blog, and she makes me feel like a lazy engineer. She keeps up a great blog, and I highly recommend that you give it a look while you’re here.

I’ve played around with each of these a pretty good bit. I run IPCop at home and at a customer site, and pfSense I’ve messed with off and on, and have been impressed by it’s performance and features.

Despite what the title of this post says, however, I don’t think that it’s really fair to compare the two head-to-head. This is because they are fundamentally different in their aims; IPCop is determined to be a good SOHO firewall, and nothing more; pfsense on the other hand, has a decidedly more corporate feel to it. In fact, in the pfsense support forums, people offer feature bounties, where someone will offer money to whomever can develop a certain feature and incorporate it into the image.  Now that is progress.

My experience in the IPcop forums has been less than stellar. I once posed a question regarding the ability to use IPcop as a transparent firewall, and was told that there would be no point in doing that with IPcop, since that isn’t what it was designed to do, and that I should try pfsense instead. Another user requested the same thing, and got the same answer. Still another user posed a question about when the 1.5 was coming out, since there were some new features coming out with it, as well as a new kernel. He was shortly ‘accused’ of being a Windows user, called a troll, and the thread locked. I read back over the posts to see if I had missed something inflammatory that he might have said to merit this, but couldn’t find anything unreasonable. That’s fanboyism for you.

I really like IPcop, and think that the core developers and the folks who have volunteered their time to create such cool addons for it have done a terrific job supporting it, but for the corporate user, pfsense seems like the better option.

It’s been a long while since I’ve posted anything here, been really busy with a lot of different projects.

The one I’m most proud of is setting up a fiber WAN connecting 5 sites @ a gig. It’s fast, really fast. I thought I would be setting it up using Procurve 3400cl switches, so imagine my surprise when Cisco 3650G switches showed up instead. No big deal really, all of the priciples apply, just the command execution is a bit different. Been up and running for close to a month now, and let me say, it is fast!

I also wrapped up a consolidate project for a customer, where I migrated four physical servers to a single virtual server. The customer has now replaced 11 physical servers with 3 rackmount ones. Virtualization isn’t a thing of the future anymore, it is now.

Got a few more projects coming up soon, I’ll post on those later.

First off, let me say that fiber optic connections using Procurve equipment do not auto-negotiate.

And yet, there is a setting to set your mini-GBIC (SFP) to 100-half, 100-full, 1000-half, 1000-full, and even auto-1000.

This is retarded. Fiber by it’s very nature is ONLY full-duplex; one strand for transmit, one for receive.

This being said I feel like you should NEVER set your fiber to anything other than ‘auto’. Sure, you can set both ends to 1000-full and it will work just fine…… what of what benefit is it? I’ve taken snapshots of interface stats, and compared between connections that are hard set to 1000-full, vs. those simply set to ‘auto’. There was NO difference whatsoever; and this was not over just a few days or weeks, but months. So then of what benefit is setting your interfaces to 1000-full? None, in my opinion. It’s just a few more lines of code that have to maintained. The KISS principle should be applied in all possible areas.

Oh and BTW, if one end of your fiber is set to 1000-full and the other to auto, you will NOT get a link.  I can’t tell you of how many new installs this has bitten me.

Your certificates need to already be in order for this to actually work.

I’ve had iManager 2.5 blown up on my iPrint server for a few weeks now, which means I haven’t been able to do any administration of my printers. That’s bad.  I tried to do an inplace upgrade to iManager 2.6, but that did nothing for me. I determined that the problem was somewhere within Apache, and ended up uninstalling Apache2, Tomcat, and iManager.

Oops.

iPrint itself is now broken. In my haste I didn’t stop to think that it actually relies on Apache for all of my printing, via the IPP URL….. so when I blew away Apache, I broke printing totally…… what now?

Time for lunch.

After eating a sub, I realized that my fix was an easy one; just reload iManager, Apache, and Tomcat off of the CD through the GUI, and it should re-write my config files. I can handle that. So I do this, and to my joy, iManager is back in business. However, printing through IPP is still not working. I check my Apache log screen, and see that it is not listening on port 631 like it should be. Now, I go digging through my httpd.conf, comparing it to my old version (yes, the backup saved my can today) and compared. They are basically identical except for one line at the bottom:

include iprint/ipp.conf

So I added this in, restarted, and everything is back online.

Now, it’s officially a ‘good day’.

Procurve switches ship with spanning-tree turned off, which is different than Cisco’s comparable switches. I have long used spanning-tree to combat network loops, which is where you have a cable connecting one port on a switch to another port on the same switch, there by causing a layer 2 packet storm.

However, this isn’t necessarily what spanning-tree is supposed to do; it is really there to provide backup links in a redundant network topology, such as a ring or a mesh. It’s ability to detect these redundant paths is what usually shuts down a network loop before it brings the whole network down.

Really though, the network loop problem is facilitated by Procurve’s ‘auto-mdix’ feature, which allows you to connect to any switchport using either a straight-through or crossover cable. Since 99.9% of networks only use straight-through cables in connecting devices to their switches, really the better solution to prevent loops is simply to disable auto-mdix on all ports (except maybe your uplinks) and leave spanning-tree off. To disable auto-mdix:

switch#(config): int a1-a24

switch#(a1-a24): mdix-mode mdix

Gotta give credit to sirmicha for the idea.