Category: Uncategorized

Jeff Nolan adds his voice to the TypePad rumblings. [here]
Take away quote:

Some very experienced IT professionals are chiming in that they understand how difficult it is to run a datacenter (something I myself have not done I would add) and I am sure the empathy is well meaning. However, there are ample examples of large datacenter operations that operate with a very high degree of uptime and performance so while running a high capacity and complex datacenter may be difficult, it is being done successfully elseware.
Mena and Ben (further putting the human face on the company, a good PR tactic) may be plenty sorry, but I’m still stuck paying for shitty Typepad performance. In fact, I’d like to see this subscription software model go even further to align with customers by putting a service level agreement (SLA) in place that lays out a series of financial penalties the provider incurs when performance and availability fall below predefined thresholds. I bet they would not have to be sorry then, because companies that have these kinds of service level agreements are always highly motivated to get it right the first time.

Wow. It’s not often you read something like this.
Oh yeah, it’s the blogosphere.
Take away quote:

The question then is: if you ran out of space and were having problems, why did you continue to take on new customers during this period? Surely a responsible business with serious capacity issues would have closed their doors to new business to assure that its current clients were taken care of.
Its called greed.

I’ve been using Ping-O-Matic for a long time, but in the last 2-3 days, it has simply not been responding.
Switched over to Pingoat — hope that works better.

Rojo has started showing unusual performance values over the last 24 hours.

CLICK IMAGE

Most of the gain has been in First Byte times — the time between the Client issuing the GET, and the Server’s response arriving back at the Client.
Since I use Rojo, this is a concern for me.

Technorati: Rojo, aggregator
IceRocket: Rojo, aggregator

Wow!

1. Engage. Participate. Be fully present. No auto-pilot.
2. Meetings and multiple appointments are a fact of work-life; the least we can do is be on time so they can start on time and our peers are not kept waiting.
3. Respect the attention of your peers. Come prepared means come prepared.
4. Always have a pen and paper for note-taking. First, you respect others who are giving you information by acknowledging it, and secondly youÂ’re expected to capture it, and follow-up; forgetting is not an option.
5. Whatever your role is, youÂ’re expected to be the expert in that role. Own it, and donÂ’t be shy about it. Stake your claim proudly. (This was part of the no bench-warmers philosophy.)
6. When you say youÂ’ll follow-up on something, do. If itÂ’s not going to happen, say so. People trip when you sweep stuff under the rug.
7. Own up to your mistakes and be okay with them. Making mistakes is perfectly fine for we all make them. However huffing and puffing about them with excuses and justifications is not fine. Get over it (we already did) and just correct it.
8. Communicate. We have found that relying on mind-reading doesnÂ’t work that well for us.
9. Trust and be trust-worthy. Much easier when Rules 1 – 8 are honored and we all keep it real.

ng this out now to be posted prominently.

For reasons that I have yet to discover, my outbound connection to the Internet appears to have been down between the times listed in the title.
As I have said before, my connection is less than “professional grade”, i.e. “don’t do this at home”.
If I find a cause, I will update.
I apologize for the incovenience.

UPDATE: Ouch.
Found a MONDO huge problem with mysqldump in MySQL 5.0.15 — it locked the tables. Totally.
I backup the database by taking a dump 4 times a day, 12 hours apart on two machines.
The Web server machine, where I was testing MySQL 5.0.15 prior to the production upgrade, is using the version mysqldump that shipped with that.
I need to read the docs and disable the table locks.

Wait! Didn’t he say he had done this already?
Yup. But forgot one stupid little step. You have to re-compile PHP to work with the new MySQL 5.0 libraries. I tried to restart Apache and it barfed all over the PHP module with MySQL library module mismatch errors…or something like that.
Done and the system is now using the right MySQL libraries.
Usual caveats about system weirdness apply. Operator weirdness is your own issue.

In case some of you want to host a GrabPERF Measurement Agent, the current prerequisites are available here.
I hope to get some folks signing up for this during this round.

Think I make this up?

These are the servers that power GrabPERF and this blog.
Don’t worry. The UPS on the floor is dead.

The TypePad application is currently experiencing performance degradation.

You could say that.

CLICK IMAGE

UPDATE: Looks like TypePad has returned to something resembling normal. Slow, but normal.

CLICK IMAGE

UPDATE: Geek News Central commiserates with TypePad. [here]

Technorati: typepad, web performance, six apart
IceRocket: typepad, web performance, six apart