I have yet to see anything appear in the S3 bucket associated with my Cloudfront distribution even though the README.txt states that, "It is safe to hit this as often as you want."

Accessing cloudfront/processqueue on my site generates this error each time:

_cloudfront_process_queue has been running for more than an hour and is most likely stuck

Comments

doublejosh’s picture

doublejosh commented

An hour seems pretty long, but it did take longer than I thought to push just a handful of images over. I imagine if I had flushed an entire preset, then generated all the images (with a generator view, command line imagecache calls, or just browsing) it would have taken some time for the transfer.

I'm also new to the cloudfront module and it's possible running the cron preprocess thing pushed the whole preset folder?!

This type of thing can sometimes go over the limits on your server's boundaries and call a memory limit error etc. which could error out your cron task.

Running re-sizes on big files, aggregating and the like have all done this to me on both dedicated servers and VPSs.

Hope that's helpful.

brianshumate’s picture

brianshumate commented
Status: Active » Closed (fixed)

OK, this error was only the symptom of another issue for which the module apparently doesn't handle the errors for.

The Cloudfront module requires that PHP has curl support, by either compiling it with --with-curl, or in the case of a modular PHP (as w/ Debian, for example) the PHP package's curl support package (php5-curl) needs to be installed.

Without curl support in my PHP, I was getting the 'cron stuck' message every time _cloudfront_process_queue() was called.

I Installed php5-curl package on my Debian systems, and now the Cloudfront module is working as expected.

I'll close this, and try to update the documentation/submit a patch for the README on this issue as well.