Archives for: September 2005


Permalink 11:52:49 am, by fumanchu Email , 193 words   English (US)
Categories: CherryPy

CherryPy 2.1 RC1 is out

The official mailing-list announcement is here.

The big change from 2.1 beta is the session filter—it's been completely rewritten.

Minor updates/fixes:

  • Server-side image maps (ISMAP) now supported.
  • More documentation in the official CherryPy book.
  • Improved coverage tool output.
  • Support for partial GET requests.
  • New HTTPError(status) exception, plus pretty HTML pages for 4xx-5xx responses.(which are customizable).
  • Separate access and error logs.
  • % HEX HEX decoding now works for URL's, not just params.
  • New cptools.serveFile function.
  • New config entries which allow you to limit the size of request headers and body (to avoid denial-of-service attacks).
  • Tracebacks can now be inserted into the CherryPy log via "server.logTracebacks" config entry (True by default).
  • New expose() function/decorator, which allows you to alias any page handler method.
  • HTTPRedirect can now be raised in _cpOnError or error filter methods.
  • Other minor bugs in the beta were fixed.

Way to go, team. I'm pleased as punch to be a part of this powerful, Pythonic product. :)


Permalink 10:04:45 pm, by fumanchu Email , 119 words   English (US)
Categories: CherryPy

mod_python wrapper for CherryPy

I wrote a WSGI wrapper for modpython a while back, but nobody's gotten it to work yet with Apache2 on Linux (at least, nobody in the CherryPy community). The current theory is that it's due to the differences in MPM's between Windows (mpm_winnt, which is one process with multiple threads) and Linux (worker, threadpool).

If you're stuck wanting CherryPy + modpython on Linux, have a look at Jamie Turner's new, which skips the WSGI layer and directly connects CherryPy to modpython. Then let everyone on #cherrypy know how it went. ;)

Permalink 05:58:38 pm, by fumanchu Email , 507 words   English (US)
Categories: IT, General

Broken mail, and when to outsource it

[Disclaimer: if you're a techie, and you find this discussion a bit patronizing, bear with me. I raise my own salary through donations, and I'm trying to communicate to my non-technical financial supporters where their money is going. If you're one of those supporters, please bear with me when I descend into meaningless jargon. ;)]

Our mail server, which runs Microsoft Exchange, died two weeks ago—critical operating system files got corrupted, and Windows wouldn't boot. This is a bad thing. Bad technically, because it meant reinstalling Windows, which was sure to confuse the Exchange service. Bad corporately, because it took us two days to get mail services back up and running, which meant lots of business not getting done. In point of fact, although we spent several hours with a consultant and a Microsoft engineer, we didn't actually fix the problem; however, we got the mail server back into enough of a working state that business could proceed. So we backed up all our mail and said to ourselves, "don't touch this machine again".

Then I started shopping. Being an intuitive fellow, I listened to that little voice in the back of my head saying, "now would be a good time to outsource our mail". That is, I went looking for someone who could run an Exchange server somewhere else on the Internet for us, and host our mail. If it worked, this would be a big benefit to us because:

  1. It would make someone else squirm when the server crashes,
  2. The provider would benefit from an economy of scale, since they perform the same service for multiple companies: they could afford redundancy and safeguards that we, a small company, cannot,
  3. It would take the administrative burden off Ryan, our system administrator. At the least, it would change his handling of issues away from technical resolutions (fixing machines) to social resolutions (calling the service provider when things go wrong),
  4. At $10/month per mailbox, it's pretty cheap for 40 users.

Would I have made the same decision if we had 20 users? 100 users? Perhaps. What was more important in my calculations was the ratio of IT workers to other staff. We are right at the point where we need more IT staff, but haven't hired them yet. That is, the demand for IT is outpacing what Ryan and I can supply.

Assume an ideal ratio of 1 IT worker for each 20 staff members. Given our staff of 40 (and growing), if we had 3 IT workers, we might have chosen to keep hosting our own mail. Similarly, if we had 20 staff to our two workers, we would have chosen (and did choose for the past few years) to host our own mail. But because our demand is exceeding our supply, we chose to outsource it for now:

Ratio of IT to staff per year

We're still in the process of moving our mail to the new provider, but don't worry about sending us mail! Nothing about that will change. I just wanted to give some of you a peek into the day-to-day (-to-week-to-month?) fires I end up putting out.

September 2005
Sun Mon Tue Wed Thu Fri Sat
 << < Current> >>
        1 2 3
4 5 6 7 8 9 10
11 12 13 14 15 16 17
18 19 20 21 22 23 24
25 26 27 28 29 30  


The requested Blog doesn't exist any more!

XML Feeds

powered by b2evolution