trouble setting up subdomain "bad_httpd_conf"

I have the domain "planetblur.org". It is being hosted on dreamhost's servers.

I have a friend who needed somethings hosted so I told him I could help him out.

I created a user "jimmyd" for him. You can SSH/SCP into the server just fine using the username.

Then, i created the subdomain "jimmy.planetblur.org". I linked that domain to the jimmyd account.

now, when you login to the server using the jimmyd account name, the folder "jimmy.planetblur.org" shows up in his home directory. I made a test index.php and put it in that folder.

now, when you browse to http://jimmy.planetblur.org, it just gives me the error:

"Site Temporarily Unavailable
We apologize for the inconvenience. Please contact the webmaster/ tech support immediately to have them rectify this.

error id: "bad_httpd_conf" "


Also, when i view the jimmy.planetblur.org domain in the control panel, there were no service records listed (A, MX, etc) like there are for all the other domains.

Anyone know how to fix this?

Edited by bculler2 on 02/23/05 11:23 AM (server time).

You probably just needed to wait for the DNS to get updated. Now it looks like you added a custom record for the jimmy domain which point to the main ip of the machine. But since the apache responsible for serving the jimmy domain has it's own ip and dns uses round-robin, half of the requests will probably work, while the other half do not.

DreamHost does not load balance customer domains so if you ever gat back two ips on a domain host lookup that is bad :
$ host jimmy.planetblur.org
jimmy.planetblur.org has address 205.196.218.50
jimmy.planetblur.org has address 205.196.219.89

Make sure you delete that custom dns A record you created or the site will always be only half working.

> Anyone know how to fix this?

Alert Support - they can temporarily clear the error condition on request.

FYI This error is a symptom of a known fault in the DH system. Support have said:

> The error occurs whenever a web or dns change is made to your domain from
> the web panel. It takes up to two hours for the server to reload and for
> the changes to go into effect. Part of the process is for the httpd conf
> file to be rewritten. There is a small amount of time when it hasn't been
> that's when the error occurs. It should clear itself up relatively
> quickly, between 15 min and 2 hours.

and

> the httpd.conf file is affected by the other domains and even though your
> domain is unchanged, you might still experience the temporary error.

In my experience it does not clear up in 2 hours. It clears up only after I alert Support.

This is terrible! I don't mind a subdomain taking some time before it starts working, but I added the subdomain w4.libertarianwiki.org, and I got this error on the domain libertarianwiki.org. The main domain is just a mirror of the subdomain. I use different subdomains to try out software before I go live by pointing the main domain to the new subdomain. Is there a work around?

I submitted a email to tech support and was shocked to find that '"bad_httpd_conf" Error' was a line item under 'Category'. Do I have to put up with down time every time I add a subdomain?

it seems that way. dreamhost seems to quickly be proving to me the "you get what you pay for" saying. we get a lot of space, a lot of bandwidth and a lot of features for the price of hosting... but the customer support, reliability and performance of things really seems to be lacking.

i sincerely hope things improve before my 90 day money-back guarantee is up.

> You probably just needed to wait for the DNS to get updated.

Surely DNS delay cannot be the cause because the browser sucessfully resolved the address. The fault occured at the resolved machine.

I've found no workaround, and would dearly like one. I've found there's a fairly high chance that any panel change to the web service config can cause this failure for that web service. Thankfully I've not experienced that same can cause this failure for all that server's web services, as Support had suggested.

Here's Support's current position, according to their last reply to me on it:

> This should not happen again, it was a one time error.

I'm waiting for clarification as to whether that means the cause is fixed.