erikdemarco

@isaumya Next time. Please please please try before you reply. ??????

I will explain why you are wrong.

CF will not will not provide you a cached response ‘IF’ there is a cookie in the ‘RESPONSE’ (IF your server response send ‘set-cookie’ in the header). ELI5: So it means When cookie exist in request, cloudflare will still respond you with cache version. Thats how all cache on earth works including nginx (CF use nginx as the backend)
source: https://support.cloudflare.com/hc/en-us/articles/200172516-Understanding-Cloudflare-s-CDN

“Cloudflare does not cache the resource if the Cache-Control header is set to private, no-store, no-cache, or max-age=0 or if there is a cookie in the response” (ELI5: CF will still cache if request header contain cookie, do you know the difference between request and response header?)
source: https://developers.cloudflare.com/cache/about/default-cache-behavior

Thats explain why using ‘default’ version work. But using ‘worker’ version doesnt work.

Have you try add to cart (when logged out) using default method instead using ‘worker’? It still send cached version. Please please please try it before reply.

After that you will understand this question. “So why it works on default version but not works on ‘worker’ version?”

“This has changed in version 1.8.0 as there will be a WordPress cron scheduled (so a scheduled event and not a true cron) that will clear the expired cache automatically.”

Isn’t the previous version works better for large site? Let’s say we set cache expiry time to 1 hour. And we have hundred thousand post that expired. Doesn’t it will gives error if all of that page not finish to be deleted in 30seconds (default php time-limit)

On previous version its more friendly to the server resources. Because it will gets deleted when the page requested. So it’s doing job one of a time. Which is very very resource friendly. Instead of doing all in one time

• This reply was modified 3 years, 7 months ago by erikdemarco.

@wfpeter Yes of course the db connection will be limited to protect server resources.

Is it possible wordfence design the the real culprit here?

Wordfence should not use sql database to do ratelimit. when site is ddosed it can be 1000req/second or more. Its way out of range for regular server to handle.
The correct way is wordfence ratelimt should run as early as possible in action queue and use file based storage. File based storage much much much more faster than db. I dont understand the logic behind wordfence team choose to store ratelimit count in sql database. Its a big flaw. It will never can handle even small volume of ddos attack.

Just by browsing all site using wordfence in here:
https://trends.builtwith.com/websitelist/Wordfence

I can make 90% of that site down just by sending very low 1000reqs/second randomized search params.

@danielbachhuber
So setting ‘WP_REDIS_DEFAULT_EXPIRE_SECONDS’ isnt enough to set the TTL?

@cusrev
Hi. I just realized cusrev already have this functionality.

How to disable all functionality and only activate this “QNA” functionality?

Just a suggestion:
Please remove this commnet meta
cr_question_reg_upvoters
cr_question_reg_downvoters

Its so useless and taking so much space if you have a lot of vote. not efficient at all serialize and deserialize it often. Alll comment system including wp karma and wpdiscuz have this slow functionality. Please consider to remove it to make the plugin more faster.

Try version WP All Import v4.6.5.

Ok Emre, thank you for clarification.
Can you recommend what is the best way to protect wpfastestcache from this kind of attac?

Hi Emre. The issue is because most cache plugin is using 1 directory to store all the cache files.

This 32000 limitation is only per-directory, not the whole system. source:
https://serverfault.com/a/46393

Is it possible to set WFC to store cache files categorized to more deep subdirectories so each directory doesnt reach 32000 files limitation?

For example the SQUID cache does it this way:
f/4b/353ac7303854033 (using md5 to make sure file count per directory not reach 32000)

Hi, I can fix it myself. But your modal name is id=”myModal”. which is very generic. you should rename it like id=”miniorange_woo_checkout_modal”

and your modal looks very bad in mobile with any theme. Looks like you just copy paste the modal template from bootstrap tutorial. Even you haven’t change the id of the modal.

Please fix it. Its very easy just search the modal html in ‘WoocommerceCheckOutForm’ class in ‘add_custom_popup’ function. and rename the modal id.

You should fix it from source. From some reason i can’t unhook this ugly modal using:
remove_action( ‘woocommerce_after_checkout_billing_form’ , array(‘WooCommerceCheckOutForm’,’add_custom_popup’), 99);

• This reply was modified 4 years, 1 month ago by erikdemarco.
• This reply was modified 4 years, 1 month ago by erikdemarco.

closed

i can confirm this is wrong solution. doesnt work.

yes thats correct. that is what i see

thank you so much, i didnt know “Test Mode” in the setting.

“we simply bundle this data with each order (along with the invoice number & date)”
Do you mean your plugin store each inovice data? So it will take up space if we have huge list of generated invoice? How to delete this data if we not require this anymore?

• This reply was modified 5 years, 2 months ago by erikdemarco.

“Am I correct in understanding that the admin will be notified for all additions to the ticket so that they can follow along and respond? Then the seller would act like an “agent” in the system, and the buyer would act like a “Support customer”?”

Yes that is correct.

Do I need to set each seller as an agent manually?
Is it possible to do it programatically using somekind of api for what i want to achieve?

• This reply was modified 5 years, 11 months ago by erikdemarco.

??

• This reply was modified 5 years, 11 months ago by erikdemarco.