My cat decided to pull the plug on my desktop system, and my /home partition suddenly couldn’t be mounted. The short story is that somehow my system was missing the jfsutils necessary to replay the journal and bring the partition up (once I realized that, the world was good).

I recently migrated from a 300 GB hard drive to a 1.5 TB hard drive, and in preparation for that, I had started backing up some of my more critical files, such as email, password database, accounting files, and so forth to dropbox via a nightly rsync. So I was able to get those things up and running on my recently-installed FC12 laptop. However, I was missing some critical emails from the day, and a perl script I had already written twice for a project that I needed to complete this morning.

The lesson learned is one that too many people fail to learn all too often until it’s too late: besides having a backup strategy, run through a disaster-recovery scenario. Yes, it’ll take you a couple of hours, but just stop and say “OK. If my desktop/server/datacenter gets hit by the digital equivalent of an asteroid, what do I have backed up? Is it sufficient? How quickly can I get back up and running with that data?”

The ultimate question is: What do I need to change?

There are some good backup scenarios out there that aren’t terribly expensive. Dropbox has a 50GB option for $9.95/mo. iDrive and several other solutions are available. Most of these have 2-4 GB free plans as well.

Ask yourself: is $9.95/mo (or higher amounts for bigger plans) all that expensive in comparison with losing all your data?

Just some food for thought…

Oh, and my changes? I moved my emails to an hourly backup and I’m going to make sure that any active project I’m working on is in version control (external server) or otherwise backed up.

I’m going to document my findings along the way, and hopefully have some insights in how to make such a trip easier the next time around

Original post here.

Check /var/qmail/queue/lock/trigger . It should look like this:

# ls -l /var/qmail/queue/lock/trigger
prw–w–w- 1 qmails qmail 0 Dec 17 21:56 /var/qmail/queue/lock/trigger

The first character “p” in the `ls -l` output means that the file is a named pipe (also called a “FIFO”). If it’s a regular file (“-”), doesn’t exist, or if the ownership/permissions are wrong, qmail-send won’t get an immediate notification when new messages are added to the queue, and it won’t process them until it hits its scheduled wake-up time, typically every 30 minutes.

It’s easy to fix:

mkfifo /var/qmail/queue/lock/trigger
chown qmails:qmail /var/qmail/queue/lock/trigger
chmod 622 /var/qmail/queue/lock/trigger

Good luck.

php_value memory_limit 32M
php_flag zend.ze1_compatibility_mode Off

As a mini-blogging framework, it is quite an interesting concept, but I don’t think it will really take off in that direction. It doesn’t give enough space to really something said (although you could shorten to txt messaging lingo, it would lose character in the translation).

What I do find is that from the standpoint of a telecommuter, it provides a sense of “inter-office chatter”. At my previous job, we would banter back and forth on random things – mostly getting a chuckle about it, sometimes getting a conversation going, but giving you a brief mental break and something away from the massively complex perl/python/shell script we were working with at the time.

For me, twitter provides that — a way to have that “chatter” without truly breaking your concentration (unless you want it to), and not be quite sucked into the inhumanity of the binary world.

Thoughts and comments welcome. You may also follow me @cybergroupdoug.

One of the staples of a programmer’s life is coffee. While I do not claim to be a coffee connoisseur, I do drink a fair amount and have a reasonable taste in what makes good quality coffee.

I recently had the great opportunity of connecting with Ken Fite on twitter, who owns a new startup in the coffee business, weeklyroast.com. I wanted to see how good his coffee was, and see if it was any more affordable than the coffee I had been getting.

In the interest of full disclosure, I offered to blog about his coffee in exchange for a free sample, which he agreed to. However, my review here is completely straight-forward and honest. Other than the free sample to try, I am not receiving any compensation for this review.

So on to the review:
I requested a sample of the Columbia Organic, which is a medium-dark roast. When I received it, it was vacuum packed in a nice heavy plastic bag, and came via USPS. The coffee certainly smelled fresh. By comparison, the bag of brand name breakfast blend that I had recently purchased didn’t really have a great aroma when I opened it – it smelled about the same as the pre-ground coffee.

I use roughly 2/3-cup of beans in a standard Mr. Coffee walmart-variety coffee grinder, ground extra fine (I find it makes the coffee smoother), for 12 cups of coffee. My coffee pot is a standard Mr. Coffee walmart-variety coffee pot (nothing fantastic there, fitting for the average kitchen).

When I brewed it up, my wife, who was suffering from a sinus infection in bed, commented that she could smell the difference from the brand name blend. While the brand name blend smells fresher than the pre-ground coffee, this was better. The coffee was most definitely fresh. My taste buds were in heaven. Now, I do buy coffee beans in the store, but Ken is right when he calls them “stale”; the flavor is not nearly as full as fresh-roasted coffee.

The price is comparable to what you would pay in the store for a pound of beans — around $8.00, plus a little bit of shipping (flat-rate). However, keep your eyes peeled: Ken has been doing giveaways and 2-for-1 deals on his tweets. In fact, he has a giveaway planned for this weekend…

Oh, and I just received a shipment of Brazil (medium-roast) coffee, my first of hopefully many orders. In case you’re wondering, I’m not planning to go back to the pre-ground canned stuff anytime soon (although, if in need of a coffee fix, you take what you can get )

ImportError: /usr/local/lib/libsvn_ra_dav-1.so.0: undefined symbol: SSL_load_error_strings

There are two possible solutions:

1. Compile and install the Neon library outside of the subversion tree, and then compile subversion against that version of neon.
2. Add the following variable to the environment of the application to force the SSL library to be loaded first:
   LD_PRELOAD=libssl.so

For a web-based application (for example, trac), add the following to the apache location directive:
SetEnv LD_PRELOAD libssl.so

Note that if using LigHTTPD or Apache/FastCGI, the environment variable may be slightly different.

The jury in Britney Spears’ trial is taking the weekend to mull whether the pop singer is guilty of driving in California without a valid license.

OK, I don’t know what it’s like in California for those who don’t have enough money to buy an extra house every other day, but here in Illinois, they wouldn’t have even walked out of the courthouse! “You’re honor, we find the little guy guilty as sin and recommend that he be locked up and have the key thrown away for the next 10 years…”

Oh well, yet another notch in the belt for stupidity and wasting people’s time.

[Edit: In case you want to read the AP article quoted above, one source is here.]

Of course, there’s always that black cat named Murphy that likes to lurk in the corners. It still didn’t work. Several calls on their part to the Data Center’s tier-2/3 teams, and still no joy. Data Center folks blaming the server, them (and me) blaming the firewall.

I finally was down to running tcpdump tonight to see just what was happening, and ready to blame the firewall for sure (as there were no packets traversing), and then, as in the words of Poirot, “the little grey cells, they have the solution!”

A quick check of netstat, and I’m banging my head on the wall. Of course. The problem was that sendmail wasn’t listening on the 10.x.x.x internal IP address!

[yoda] The Little Things – Those you must beware! [/yoda]