Prx

Hello,

Yes, I mean the “Full Site Delivery” option advertised in the CDN module of W3TC. Could it be an alternative to AWS CloudFront?

ph.

Hi,

This seems to help: I ran some test, loading media-heavy entries, and the performance is much better I think. I tried using test browsers (no extension, no caching), and also using a VPN to access the site from different points. I’ll keep monitoring it.

If I may ask, how FSD compared to CloudFront?

Update:

? I deactivated some plugins, and revert back to 2.2.4.
? Now it can pass the distribution test
? However, if I enable the CDN (and purge the cache as usual), I’m still seeing slow loading time for media content, along with 504 error.

Heloo,

I went on and manually created a new distribution in AWS CloudFront, using the same parametres.

? Same origin: that would be my website example.com
? Viewer protocol policy: HTTP or HTTPS

In W3TC cache CDN dashboard, I then replaced the new Distribution domain name in the field “Replace site’s hostname with:”

“SSL support” is set to “auto”

This result in the same failure: the cloudfront distribution fails. Same if I try (again) to create an origin. To be on the safe side, I saved the new info and purge the cache before running theses tests.

Thanks for your ongoing support!

p.

Hello,

I reverted to 2.2.3 (using the package available here), emptied the cache, but the test still fails.

To revert, suffice to download the zip file of a previous version and install it through WordPress dashboard (without touching the installed version). There will be a prompt asking if the version being installed should replace the currently installed version. That’s how I reverted back.

I can run more tests.

Thank you Marko: I will check. I highly appreciate the prompt reply and the help!

Hi,

1) I’m using Amazon CloudFront. I went back in W3TC setting and re-created the distributin and it seems to have fixed the issue (previously, I had also “tested” the CloudFront distribution and it had passd the test;

2) I purged the CDN because after the recent updated my plugin for social buttons couldn’t update (Social Warfare, I’m in contact with them as well). It seems that when the CDN is disabled, all is fine, but if I enable it, the social buttons do not displayed properly (after the update; all was fine before).

3) After the update (that is WP update, followed by an updated for all my plugins) I first cleared the cache, and then purged the CDN. Is there something else I need to clean/purge?

But in any case, CDN is working and W3TC is communicating properly with my Amazon CloudFront account.

Thanks!

Thanks a lot. Since I already have my S3 buckets setup with W3 Total Cache, am I correct to assume I can skip to step 4 of the installation process for the S3 Image Optimizer plugin? I’ll check first to see if the images that I have already optimized were updated in the cache.

P.

Hi,

Thanks a lot for the quick reply!

1) I’ve already wiped out WordPress and tried again 4 times. Each time I do a clean install, deleting all WP files and re-installing from scratch. I also delete the MySQL database and create a brand new one.

2) 93 posts.

3) Error logs display the following errors (I have anonymized the URLs):

[Sat Mar 21 09:10:59 2015] [emerg] [client 00.00.00.00] (22)Invalid argument: mod_fcgid: can't lock process table in pid 30445, referer: https://mywebsite.org/wp-admin/
[Sat Mar 21 09:11:02 2015] [emerg] [client 00.00.00.00] (43)Identifier removed: mod_fcgid: can't lock process table in pid 10277, referer: https://mywebsite.org/wp-admin/
[Sat Mar 21 09:17:28 2015] [error] [client 00.00.00.00] File does not exist: /home/adminname/mywebsite.org/internal_error.html, referer: https://mywebsite.org/wp-admin/admin.php?import=wordpress&step=1&_wpnonce=3c32d54962
[Sat Mar 21 09:21:07 2015] [error] [client 00.00.00.00] File does not exist: /home/adminname/mywebsite.org/favicon.ico
[Sat Mar 21 09:21:07 2015] [error] [client 00.00.00.00] File does not exist: /home/adminname/mywebsite.org/missing.html

Hi,

So what I get when I try to import is either a hang (it will freeze or hang for ever) and an “Internal Server Error”.

I have activated the “debug” mode in the importer. After a hang, if a refresh the page, here is what I get:
Failed to create new user for nameofuser. Their posts will be attributed to the current user. Sorry, that username already exists!
Page “NameofPage01” already exists.
Media “imagefile.jpg” already exists.
Media “NameofMedia01” already exists.
Media “NameofMedia02” already exists.
Page “NameofPage02” already exists.
Page “NameofPage03” already exists.
Etc.

And indeed everything was imported, except every single post has lost its original author and has instead been attributed to the administrator in the process.

P.

Thank you for your feedback. I’ll put more information next time. I’m using Origin Pull with Amazon CloudFront.

But it turned out that the problem was completely unrelated to W3TC. It had to do with my coding. So this one is resolved.

A couple more information about debugging this problem:

– A search with the “site” operator will turn out the post. That is: if I ask Google to search for some keywords not in the whole wide web, but specifically in my site, my post will turn up in the results.

– I use the Structured Data Testing Tool to examine how Google was reading my post. The only strange thing is that the information for the hcard belonged to a blog that had reblogged my post: it appeared as a trackback under my own post.

– It’s been nearly a week since I published this post, and it still won’t show in the results (but links to its tags, category, etc. do show).

– Now it’s simpler for me to resubmit the post via Google Webmaster Tools (10 such submission are allowed each month).

Thanks.

That’s interesting. Thank you for your input.

(For others interested in this thread: enabling access log in CloudFront can be done when creating the distribution. If the distribution is already created, logging can be enabled by editing the distribution options. If one uses a custom origin (say for origin pull distribution), then one will need to provide a valid S3 bucket for the logs to be dumped in.)

Hi Daniel,

First: Happy New Year and best wishes for 2014.

I was happy to see the update. I have installed it. Things run a bit smoother. No more errors, but I still get a “WARNING” when the backup is large (for example: wp-content).

I have already split by backup workflow in multiple events. What can I do next?

See log below:

[28-Dec-2013 06:00:51] 1. Trying to create backup archive …
[28-Dec-2013 06:00:51] Compressing files as TarGz. Please be patient, this may take a moment.
[28-Dec-2013 06:03:38] Backup archive created.
[28-Dec-2013 06:03:38] Archive size is 667.96 MB.
[28-Dec-2013 06:03:38] 875 Files with 671.55 MB in Archive.
[28-Dec-2013 06:03:38] 1. Trying to send backup file to S3 Service …
[28-Dec-2013 06:03:38] Connected to S3 Bucket "mywebsite" in
[28-Dec-2013 06:03:38] Checking for not aborted multipart Uploads …
[28-Dec-2013 06:03:38] Starting upload to S3 Service …
[28-Dec-2013 06:15:42] WARNING: Job restarts due to inactivity for more than 5 minutes.
[28-Dec-2013 06:15:42] 2. Trying to send backup file to S3 Service …
[28-Dec-2013 06:15:42] Connected to S3 Bucket "mywebsite" in
[28-Dec-2013 06:15:42] Checking for not aborted multipart Uploads …
[28-Dec-2013 06:15:43] Starting upload to S3 Service …
[28-Dec-2013 06:15:49] Backup transferred to https://s3.amazonaws.com/mywebsite/mywebsite/backwpup_0833b1_2013-12-28_06-00-51.tar.gz.
[28-Dec-2013 06:15:49] WARNING: Job finished with warnings in 898 seconds. Please resolve them for correct execution.

Hi Daniel,

Thanks for answering. I have take good notice: new version tomorrow, I will make more test.

As for the email here is what I have:

1) I have the correct email address in the field “Send log to e-mail address”

2) In “E-Mail FROM field” I have simply put “BackWPup”

3) In “Errors only” I have CHECKED ” Send e-mail with log only when errors occur during job execution”

This job ran last night, there was 1 error, two warning, and I didn’t get any email.

4) I know WordPress can send me emails because I get one whenever my blog has a comment or a pingback;

5) I have checked my spam folder and there’s no email from BackWPup in it.

Hope it clarifies things a bit.

P.