Mercury

I have setup a few sites using the Pantheon Linode stack script (what a blessing!). On one of them, I want to totally disable Varnish, but keep all other settings on the server (memcache, etc). If I kill all Varnish processes, the site goes down. How to get back to good ol' Apache only?

To Josh and Greg,

I know Mercury 1.2 is still in development, so I thought you'd want some error feedback. I've gone to launchpad and followed the following instructions:

<

blockquote>To get up and running on Trunk:

Start with a fresh installation of Ubuntu 10.04LTS
apt-get install bzr fabric
bzr branch lp:pantheon /opt/pantheon
cd /opt/pantheon/fab

The last command should one of the following:
fab initialize:vps=aws (for instance-based AWS servers)
fab initialize:vps=ebs (for ebs-based AWS servers - experimental)

crontab is failing to run with this error

Fatal error: Call to a member function increment() on a non-object in
...../sites/all/modules/memcache/memcache.inc on
line 262

any ideas for me please ?

Regards Allan

We have a mercury server setup on VPS.net that is using multi-site to run multiple sites using the same Pressflow install. We recently ran into a problem with implementing our non-www to www canonical domain redirects. In the past we were able to uncomment the section of the .htaccess file included by Drupal to do these redirects because we were not using multi-site. Now that we are using multi-site we have tried to implement the redirects using apache vhosts but the redirect never seems to happen and it appears to be some kind of interference from Varnish that is keeping it from working.

Is it possible to use AWS RDS? If so, how would that affect performance? (An estimate and/or explanation of the MySQL optimizations that are affected would be appreciated.)

Thanks!

This has been an ongoing problem for my site that I have still not figured out.

I am running Mercury but my site can barely keep up with just 20 users. The site always comes to a halt. Even with just 1 or 2 users the site is very slow. I am not even sure where to begin diagnosing the issue as I am not much of a server admin. I know I have some slow database queries but even on pages with no major queries the site just hangs for a long time. Im getting really frustrated because my shared server was 100x faster then my current Mercury setup with the same queries and everything.

Firstly, many, many thanks for all the incredible work on Pantheon. Seriously cool stuff.

I've been trying to install Mercury 1.1 on Rackspace and got as far as Hudson running happily. When I try to build 'mercury_init', Hudson comes up with errors during the postfix setup - relevant chunk of console output below.

Lengthy googling of whiptail and terminal.debconf has only revealed a whole load of very old and unresolved bugs concerning whiptail's manual use and its incompatibility with some terminals, though my google-fu may not have been at its best.

Any ideas?

Working on getting a Multi-Domain, but not multi-site install on Ubuntu 10.04 working.

The reason for not having it be a Multi-site install is that we deal a lot with Japan and Japanese clients. Core does have to be patched from time to time. Now if I setup a server that has two different Drupal sites going, one might need core patched a bit, and I don't want all the sites to have the burden of having that patched install.

I will post back with my workings, but if anyone could post some tips that would be great too.

We’re running into an issue sending duplicate notifications to our users using the Notifications module on our Mercury Pressflow implementation. The duplicate messages are identical save one thing- the [node-url] token is being replaced with ‘default’ in one of the messages. All the other tokens in the message are being replaced correctly.
The duplicate emails do not happen consistently, maybe 10-15% of the notifications sent out, however a duplicate message always has the proper url & the ‘default’ url.