There’s 2 big problems with that:
1) This is a live site promoting and event and selling tickets for an NYE party, there’s no way I can switch themes and deactivate a ton of things then wait for transactions to go through.
2) It happened on two different sites hosted on the same 2 server wordpress cluster on the same day, and each only missing a single transaction. Neither has done it again since then.
I was hoping there might be some way of turning on some more verbose logging and just let it run in the background so I can capture more relevant info if it happens again.
The only log that has anything in it is the “woocommerce-gatewat-stripe-<someuuid>.log” and it doesn’t have anything about the orders that weren’t charged. There’s no entry for them, it’s as if the system didn’t even try to charge the card. There’s not stripe failure, no error messages, nothing. The order before that one (397) has a full log entry where everything works, the one after is the same (409), but the order that didn’t get charged (403) doesn’t appear in the logs at all.
I’ve got another example here….
Right now if you go to my front page https://www.wizardweaving.com, I have the featured products shown under some text. For some reason, one product which I just added shows up square as the 300×300 cropped version. As seen in this screen shot: (with developer tools showing its calling a different image than all the rest.)
But then if you click the menu in the header under “Shop All Products” and go to “Apple Watch Bands” Those call come up called with the proper 300×300 thumbnail:
This is what I’m talking about… and it’s a major hassle because it cause my front page load to be nearly 5MB when really, it should be less than 1MB just like this product page.
THIS is the problem I need to solve.
I have tried without the CDN yes.. the CDN is just a passthrough, so whatever the page requests will be pulled through the CDN instead of the regular site. The problem is, if you look at the page source, the page is looking for the full size, full resolution image, not the thumbnails that are generated for catalog and product previews.
Yeah, Varnish is a little weird for sure, but yes… I am running version 4. Version 4.0.3 to be exact on Ubuntu 14.04 with nginx 1.10.1 in front of it to strip off SSL and haproxy behind it to load balance between my two wordpress and DB nodes.
As for the IPs, I think they’re correct because if I put in IPs that DON’T match I do exactly get an error and it tells me it can’t be done and with the right IPs in there, from the varnish log I see the “matched localhost” entry… which is odd, since it’s not the localhost, but it does work. Including when I edited a single page.
If I run the curl command from either of my wordpress nodes I get the following. I’m running “varnishlog | grep -i purge” to minimize the huge output, but in response I get:
- ReqMethod PURGE
- VCL_acl MATCH purge "localhost"
- VCL_return purge
- VCL_call PURGE
- Debug "VCL_error(200, Purged)"
- RespReason Purged
On the wordpress node I ran the curl from, I get:
<!DOCTYPE html>
<html>
<head>
<title>200 Purged</title>
</head>
<body>
<h1>Error 200 Purged</h1>
<p>Purged</p>
<h3>Guru Meditation:</h3>
<p>XID: 87</p>
<hr>
<p>Varnish cache server</p>
</body>
</html>
So that would seem to me like it works, no? But then, when I just checked the homepage of my site, the initial request shows in the response headers that the page is still has Age: 2770 in it which seems to indicate that varnish in fact did NOT purge the whole site.
