In my last post I discussed using HAST with ZFS. That tells you how to replicate a filesystem back and forth between two machines. That’s nice, as far as it goes, but I want automatic failover. Clustering. I want to wake up in the morning to a message that says “machine 1 failed, machine 2 took over, and nobody noticed” instead of a lot of messages from angry customers. The standard FreeBSD failover mechanism is CARP, the Common Access Redundancy Protocol. Here’s the basics of CARP. Continue reading “CARP and devd on FreeBSD”
There’s a nice tutorial on using HAST (Highly Available STorage) with UFS and ucarp. That’s very nice, but in my failover scenario I can’t use UFS; a fsck would take too long, and a background fsck would be most likely to lose the data I’m most likely to need. And FreeBSD comes with a kernel-side CARP implementation; why would I use the userland implementation instead? So: the tutorial is great, except it doesn’t do what I want. I’ll attack this problem in two phases: one, get HAST with ZFS running, and experiment with it. Two, get CARP failover to trigger HAST failover automatically. (I believe I can use devd for CARP-initiated failover, but I’ll need to do further research on that. That’ll be another posting.) Today I’m experimenting with HAST and ZFS. Continue reading “HAST and ZFS”
Daisuke Aoyama (Google translation) has created a kernel module to glue an iSCSI disk to a boot drive. While the driver was intended for use with iBFT, you can also use it with gpxeboot. He has even made FreeBSD install ISOs with iSCSI support available. The server has limited bandwidth, so start the download well before you want to use it. (I have previously looked at using iSCSI disks on FreeBSD.)
Thank you, Aoyama-san!
I’m making an effort to work in the same way as my co-workers. This means using a Windows laptop, after fifteen years of Unixish desktops. I like to change desktop operating systems every couple of years anyway, so this isn’t a huge deal. The new work laptop came with Windows Vista, HP Bastardized Overloaded Nagware Edition, so the company Supreme Leader got me a Windows 7 DVD and license. I threw the disk into the laptop, kept hitting ENTER until the OS was installed, fed it my license key, and was up and running. That almost destroyed my productivity forever.
In the time between when a book is released and when the first reviews appear, authors experience a variety of symptoms ranging from unease to outright panic. “Did I miss something obvious?” “Did I really screw up?” “Am I the laughingstock of the tech industry?” Web pages can be fixed, but printed books are immutable; any errors are immortalized. Over the years, I’ve gotten the symptoms down to an ongoing gentle nausea.
Thankfully, the first two reviews on Network Flow Analysis are out.
Then there’s a review from Dad of Divas. DofD reviews lots of stuff on his blog, mostly stuff that he receives as a promotion. The odd thing is, I have no idea how he got a copy of the book. The lesson? The publishing publicity machine is mysterious and inexplicable, even to those of us trapped inside it.
But, having read the reviews, my lingering dread and nausea is gone. Thanks, folks.
We got a new Synaccess NetBooter networked power switch. Not only does this device do SNMP, but it supposedly reports on temperature and power utilization via SNMP. These are useful things to alarm on, and even to graph. So, I’m creating Cacti templates for them. Continue reading “my fourth Cacti template”
What, it’s been more than a week since I posted? That’s easily explained: I’m not doing anything interesting.
Don’t get me wrong, I’m busy. Project Albatross (my current personal writing project) is coming along nicely, at 25,000 words out of approximately 80,000. But blogging about work that is unlikely to be published is downright narcissistic. For those of you waiting for me to pump out a new technical book, I can tell you that my goal is to have PA complete before World Fantasy 2010, so I have two projects to sell instead of one, thus doubling my chances of further prostituting myself to the media industry getting an agent for my non-nonfiction and selling same to a publisher. The new nonfiction should start rolling shortly thereafter.
In the day job, I’m debugging Cacti 0.8.7f. Not interesting.
I’m also configuring OpenNebula. Not interesting.
Installing Drupal? Not interesting.
I’m planning to work on an interesting FreeBSD-based project soon, but I’m not allowed to discuss it in public. When I can, when I finish, if I finish, I will. But today, not interesting.
Debugging OpenSolaris NFS? Most of my debugging hints are taken from other people’s blogs. I’m keeping notes, mind you, and in the event that I learn something interesting, I’ll post it. But mostly: not interesting.
LDAP? Surely I have some LDAP pain to share? Nope. Once you get through the initial torment, expanding LDAP is rote. Not interesting.
Personal life? I’m not going to blog about that, unless something truly spectacular happens.
My life is basically boring. Your attention and time are valuable. I wish to not waste it. If you’re spending your time reading this blog, chances are I value your time more highly than you do. I highly recommend using a RSS reader, so you’ll get notified when I have a new post.
But, in an effort to provide useful content this week, I highly recommend Janet Fitch’s article on 10 Writing Tips That Can Help Almost Anyone. Most of these “10 tips” lists are as a waste of space, but this particular one isn’t. If you write for any purpose, read it. If you expect me to read your work, follow it.
Here’s hoping I do something interesting soon.
I’m back from Toronto. Now that I’m caught up on deleting spam, I’ve been able to get to the post office. All of the free review books are now in the hands of the post office.
As I write this, somehow Network Flow Analysis is Amazon’s #4 best-seller in the category “Production, Operation, and Management.” Amazon’s classification algorithms appear to be smoking some good stuff, but hey, it’s a best-seller list, so I’ll take it.
I’ve gone through the comments on my offer to give away review copies of my new book, and to my surprise found that several people I’d like to give books won’t be getting them… because I can’t contact them.
Leaving a comment that presents you as an intelligent, capable reviewer is great. I appreciate that. But if I can’t contact you, if I can’t write you and ask for your address, then I can’t send you a book. Some people left Web addresses. Fine, I can look there and get your contact information… if it’s on the page, or if you have a contact form, or something. A few people have nice blogs, but no apparent contact form. If your blog is not in English, and didn’t have something that looked like a contact or an email address, I looked at the front page source code to find a mailto: link. That worked in some cases.
I’ve emailed people I can email, to get their snail mail addresses, so I can ship the review copies I have. If you posted a thoughtful, insightful comment, but didn’t leave me a way to contact you… sorry. Books will be shipped out next Tuesday or Wednesday, after my Toronto trip. And if you ever get to Toronto, check out Mysteriously Yours.
Rancid, or a similar program, is a necessity in network management. You can find a decent rancid tutorial at http://www.joe-ma.co.za/page.php?9, so I’m not going to flog that dead horse. I can’t expect my coworkers to learn either CVS or Subversion, however. They want a pretty Web interface, or they won’t use the tool. Continue reading “rancid and cvsweb”