More GoDaddy Woes

My GoDaddy Rails site has been plugging away for a couple of weeks.
Its just a proof of concept with very little use. I haven’t updated
it or restarted FCGI at all. But today, it just quit working
altogether. In the error log, I’m getting:

(104)Connection reset by peer: FastCGI: comm with server
“dispatch.fcgi” aborted: read failed
FastCGI: incomplete headers (0 bytes) received from server
“dispatch.fcgi”

I did some searching on these, and most of the suggested cures were
to reinstall FCGI. I deleted the session files, restarted the FCGI
processes, but nothing changed the behavior.

At this point, I’m wondering if maybe its just my particular server
that’s hosed, or if GoDaddy has upgraded something in the environment
that might be affecting Rails and FCGI in general - anyone else using
GoDaddy seeing a sudden onslaught of weird behavior around Wed
afternoon? It’s starting to look like this is a cheap and easy
solution, but not necessarily workable even for an app that you just
want to demo now and then . . .

SR

Steven,

I completely feel your pain. Over the past month I’ve struggled with
GoDaddy time and time again. I was helping a friend with an app that we
got running two weeks ago and last week it stopped working out of
nowhere. There haven’t been any code changes and I started seeing this
message in the production log:

product.rb:2: parse error, unexpected tCONSTANT, expecting $
vti_timelastmodified:TR01 Aug 2006 17:07:07 -0000

So I opened up a ticket and got the following response from GoDaddy
support:

“Thank you for contacting customer support. Unfortunately we are unable
to provide third party scripting assistance via this forum. We apologize
for any inconvenience this may cause.”

No insight to what was causing the problem, just a friendly “it’s your
problem not ours” answer. Like I said this came out of nowhere and I’m
now investigating other Rails hosting services.

I too think something has changed with GoDaddy’s hosting environment,
but am to the point where I’m ready to switch. They’ve caused me too
much time and frustration and my app performance is horrible! Even
running FastCGI I’ve run into 10-15 second response times for basic
requests.

At any rate, I just wanted to let you know that you’re not alone and
needed to vent a little :slight_smile:

Good luck,
Dennis

Just to add to the go-daddy bashing, although this isn’t rails
related…

I had a dedicated server, in production, with GoDaddy. One day, the
server
started going down for no apparent reason. Keep in mind, this was in
production with a few sites that have reasonable traffic.

Several calls to support and the same response “please put in a reboot
request”. The request can only be scheduled in 15 minute intervals. As
soon
as it would come back, it would go down within 2-3 minutes. Then I had
to
wait another 10 minutes for a reboot request.

We kept calling were eventually told we were overloading our server (we
had
3 web sites and about 100 email boxes) and basically it wasn’t their
problem
because they can’t support anything that we are running that is causing
it
to go down.

After some deep investigating throughout the evening (during the brief
periods between reboot requests), I found the core-temp had been getting
pretty high generally right before the server would lock up. After
telling
go-daddy this, they finally decided to investigate. They found there was
a
bad fan on our server and it was causing the system to overheat and
shutdown.

Needless to say, we were really pi$$ed about this. For the first few
hours
of the issue, GoDaddy wouldn’t even acknowledge the problem could be on
their end, and wouldn’t even investigate. They had kept telling us they
were
looking into it. We were watching “who” on the server, and nobody had
ever
logged in. Now, of course, hardware fails all the time, I can’t be angry
about that— but the level of service and responsiveness by GoDaddy was
horrible… as if it really didn’t even matter to them (not that I
suppose
it does).

The best is that while we were waiting for a response from level1-tech
who
was waiting on a response from level2-tech, he said “You know, I see
here
your account has a few domains up for renewal, would you like to
consider
renewing for 5-10 years?”. He then asked if we would like to switch our
server plan from month-to-month to a yearly contract. I just laughed.

When the fan finally got switched, we got a new server with rackspace
and
cancelled the godaddy one a couple days later. I even transferred my
domains
off their registrar (paying transfer fees to the new registrar) just out
of
spite.

Dennis B. wrote:

I completely feel your pain. Over the past
month I’ve struggled with GoDaddy time and
time again.

I’m now investigating other Rails hosting services.

I’ve been using a2hosting for the past month or so with very good
results.
I recommend you check them out.

Best regards,
Bill

Bill,

Thanks much, I will check them out. As a follow-up to my previous email
I was able to finally get the app working after hacking a bit and
finding this in the error log:

“Cannot find gem for Rails =1.1.3:
Install the missing gem with ‘gem install -v=1.1.3 rails’, or
change environment.rb to define RAILS_GEM_VERSION with your desired
version.”

Supposedly GoDaddy only supports 1.0.0. Yet changing RAILS_GEM_VERSION
in environment.rb to 1.0.0 still caused problems:

“Cannot find gem for Rails =1.0.0:
Install the missing gem with ‘gem install -v=1.0.0 rails’, or
change environment.rb to define RAILS_GEM_VERSION with your desired
version.”

I decided to comment out that line completely as a last resort and the
app started working. Who knows, it will probably break next week, but
at least it will buy me some time until I make the switch.

Thanks again,
Dennis