* disable warning “CLOSE-read” (fixes #2091)
* mod_rrdtool: fix creating file if it doesn’t exist (#1788)
* mod_fastcgi: fix mod_fastcgi packet parsing
* mod_fastcgi: Don’t reconnect after connect() succeeded (fixes #2096)1.4.25 - the slogan is a lie - Lighttpd - fly light
Anyway, the following things have changed since 2.2.5: Fixed compatibility with /tmp cleaners such as tmpwatch Some /tmp cleaner programs such as tmpwatch try to remove subdirectories in /tmp/passenger.xxx after a while because they think those subdirectories are unused. This could cause Phusion Passenger to malfunction, requiring a web server restart. Measures have now been taken to prevent those tmp cleaner programs from removing anything in /tmp/passenger.xxx. Issue #365. Autodetection now gives more priority to Rack than to Rails When autodetecting the application type, Rack is now given more priority than Rails. This allows one to drop a config.ru file in a Rails directory and have it detected as a Rack application instead of a Rails application. Patch contributed by Sam Pohlenz: issue #338. Increased default socket backlog The default socket backlog has been increased from ‘SOMAXCONN’ (which is 128 on most platforms) to 1024. This should fix most ‘helper_server.sock failed: Resource temporarily unavailable’ errors. Apache fixes The location of the ‘apxs’ and ‘apr-config’ commands can now also be passed to the installer through the –apxs-path and –apr-config-path parameters, in addition to the $APXS2 and $APR_CONFIG environment variables. Issue #3. Nginx fixes and changes Various problems that only occur on 64-bit platforms (such as Snow Leopard) have been fixed. The installer now installs Nginx 0.7.64 by default. Various other fixes Fixed compilation problems on Solaris. Issue #369 and issue #379. Fixed crashes on PowerPC. Some Ruby 1.9 compatibility fixes. Issue #398. The installer now displays correct dependency installation instructions for Mandriva Linux.
9:07 AM PST This issue is localized to a very small number of instances in a single Availability Zone in the EU-WEST-1 region. We have identified the cause of the problem and are working on fixing the affected instances. Customers experiencing DNS issues from their instances can chose to relaunch their instances if they require an immediate fix.