Mikey

@qtwrk, perfect, that’s exactly what I wanted.

Thank you.

Hi @qtwrk,

Thank you for the response, but I’m afraid you misunderstood my question.

So, let’s say I have a form on frontend. When the user submits the form, a php function is executed. Now, I want to insert inside that function a trigger for purging all pages (Note: not purge all, but only the pages).

Basically I want to call function like Litespeed\Purge::purge_pages() inside the function for the submission of the form.

Hope it’s clearer now.

Thank you.

Ok, this is weird, I just noticed that it only happens when using Mozilla browser.

The files are not requested twice on Chrome or Opera, for instance.

Well, it must be something caused by LSCache since it only happens when the plugin is enabled.

Did you get a chance to run a test and check if you also have this issue?

Because in this case I will have to check the LSCache files to see what initiates it, but I was expecting it to be a general issue with the plugin.

I thought you said I saw it from root domain in your previous comment. So you don’t see my request at all?

I’ve just joined Slack (username is same as here – mikeyhash), however I don’t know how to use it (yet).

Yes, I’ve made a request from the root domain (which is the live website) – https://prntscr.com/vpgne4 (in case you’re wondering, I use screenshots because I don’t want this forum page to appear on Google searches).

And it doesn’t work, the status remains at 3 (requested) and the image cannot be pulled.

What do you see in the logs? What’s the issue?

@qtwrk, I believe my last requests were made more than 48h ago.

I just sent another request right now from the main site domain. Please have a look.

@qtwrk,

https://prntscr.com/voa6oa – first stage subdomain, where Image Optimization WORKS
https://prntscr.com/voa76n – second stage subdomain, where IO did not work

On the main site domain, IO acts the same as in case of the second stage.

Yes, I understand that accessing REST may sometimes be tricky, but then again, how come there are no issues for one subdomain (we’re basically talking about the same site, server and environment as the others)?

As regards your security question, I only use Wordfence, but:

1) On my first stage, IO works with Wordfence enabled and without even being necessary to whitelist your cloud servers IPs.
2) I’ve run tests with the IPs whitelisted.
3) I’ve run tests with Wordfence completely removed.
4) The .htaccess file does not contain any rewrite rule concerning security.

I look forward to something from the image server logs that could help solving this.

Many thanks.

@qtwrk, well:

1. I’ve tried destroying all optimized data, but I’ve also tested the latest version directly on a fresh environment, so the database structure failure is out of question.

2. This is also out of question, since REST API is allowed (mydomain.com/?rest_route=/ is all fine and my tests are without any security plugin, firewall or CDN service.

Moreover, if any of this common causes were present, the image optimization would not work on all of my subdomains (but, as I mentioned above, I have that first staging site where the optimization works well).

This is getting really annoying :).

Any other possible causes (irrespective of how weird or improbable they might seem)?

@qtwrk, thank you for your answer.

My issue is the same as OP’s ( @overflame ), so it would rather be a duplicate topic.

I only pointed out the tests that I’ve run on the issue, in order to try to figure out what the cause may be.

Any possible causes that you might think of for such behaviour?

Many thanks.

@qtwrk, I’ve just tested again on a new staging site.

I first removed from it any trace of LSC plugin (all tables and all related options from wp_options) and freshly installed the latest version from github (i.e. 3.6-rc12).

Now, I have the live site, a first staging site (where I first tested LSC and the image optimization works just fine) and the above-mentioned new staging site, basically all sharing the same IP address.

Is it possible that the QUIC server only allows one connection per IP address even though it doesn’t show any errors and even though there is a different Domain Key for each site?

As I mentioned, on the first staging site where I first installed and tested the Image Optimization, it worked and continues to work without any issue. For any other sites (the live one and subsequent staging), the image optimization does not work, as in the status never updates to 6 (notified) and thus the images cannot be pulled, as per my previous comment.

I have the exact same issue. The optm_status in the database tables remains at 3 (requested) and thus the “Pull images” button will not become available.

Here’s a screenshot: https://prntscr.com/vjkw1d. It seems that the images are pushed to the Cloud server and accepted, but that’s it, the status never updates to 6 (notified).

HOWEVER, please note that I’ve created a stage site (basically a pure replica of the live one) and everything works well on the stage site. I have no idea how this could be possible, since it’s basically the same site, server and environment.

I’m using the latest version from github.

Any suggestions, @qtwrk?

• This reply was modified 4 years, 4 months ago by Mikey.

@alphaexcapital

How did you rollback to pre-3?

LS Cache plugin doesn’t offer a “rollback” option, so I had to take care of it on my own, since it involves operations in the database in order to clean things up and revert to the previous litespeed_img_optm table structure.

Now, I’m a php developer and familiar with sql, thus I could safely do it properly, but if this is not the same case for you, I’m afraid you’ll just have to wait for the LiteSpeed guys to solve these issues in 3+.

Hi @qtwrk,

About mobile issue , have you enabled mobile cache ?
I noticed in some sites , mobile cache is not working properly unless
cache mobile is enabled , despite the site has responsive theme.

Shouldn’t this be further investigated and taken care of? I have never enabled mobile cache, since my site has always been fully mobile responsive, but I will run some tests on a stage site with v 3+ and revert with findings.

Hi @alphaexcapital, thanks for reporting this.

Plus the website does not always load nicely through Facebook Mobile (It did pre-3+ version).
I actually can’t believe I haven’t tested the mobile Facebook and Instagram browsers, since those are really important nowadays. I guess not working properly on Safari was enough for me to immediately revert to v. 2.9. I can imagine that on Fb/Ig browsers is worse.

The other thing is the images don’t render – they have the placeholder and that is it, even with Lazyload on – nothing is triggered to activate the images.
I personally haven’t actually noticed issues with the Lazy Loading (I also serve .webp images), but I will try to further test this as well.

@hailite, yes, more clear now.

Well, yes, as I said, saving stuff in postmeta is the easiest way to go ??

But I would say that the delete_post hook can be used very easy, together with the simple function for the sql deletion from the new custom table.

Perhaps you’d reconsider this.

In the meantime, I will probably change the code on my website so that it doesn’t save the litespeed-optimize-size postmeta anymore.