Search results

Hi, lscache writes files out as: <webserver-user>.<user-group> e.g. nobody.myuser Unfortunately on Plesk this falls apart as Plesk doesn't create a group per user, so you end up with: apache.psacln This is problematic as it means the files in the lscache directory don't count towards a...

Hi, We're seeing an issue (tested on several of our installed) where if a request sets an authorization header as such: Authorization: Apikey <key> This does not get set in PHP and HTTP_AUTHORIZATION is not present. The same happens with Digest as well. If the request uses Basic or Bearer...

The dreaded: net::ERR_QUIC_PROTOCOL_ERROR This occurs sometimes on 1st page load of a site, a number of the images will fail to load. Looking in developer tools in Chrome we see the quic error. Which seemingly results in the full image data not being sent in response: Response header says...

Not sure this is specific to Litespeed. We sometime see this issue, cPanel ends up doing an orderly shutdown of Litespeed. We also saw it the other night, but we also saw it with servers still running Apache as well.

Yes, I've checked on several, a 206 response is returned on the first request the image not in the cache with a content-range. All clients then get a broken image as the partial image is cached and returned to them.

Unfortunately I can't answer that exactly, I can see 206 responses in the logs. But I can't tell you if the original request that generated the cached image was a 206 response or not.

Okay, that's not a problem, we can do that - but surely it's a bug if LS is returning partial content for a client that isn't making any (let alone the same) content-range request? Yes, we're using a RewriteRule with [P,L,E=Cache-Control:max-age=3600]

We'll beg to differ on that :) x-litespeed-cache:hit Not a lot I can tell you other than Litespeed Cache is set to cache to a RAM disk, only cache objects up to 128k in size, in the public cache, and cache requests with query strings. These images are generated dynamically, so they are...

Hi, Since 5.1.13 (and in 5.1.14) we're seeing Litespeed occasionally serve broken images. We've tracked this down to it serving a partial content range to clients - I am guessing this is because the initial requestor of the image, called it in multiple parts for whatever reason and the last...

Hi, They were originally made without in, my test script makes them without it (our script purges Litespeed and Varnish, or whichever the user chooses). Also these commands are already been run from the server. I'd even added in the servers public IP as trusted and 127.0.0.1T as well before I...

Uhmm.... because then the request just 404s. Initially we didn't have a RewriteCond at all, but it made no difference. We added it in as a "just in case" because it is what it says in the docs to see if we could get to the bottom of why it wasn't purging. I really can't believe a product we...

User <-> Litespeed (with lscache enabled) acting as a proxy <-> Varnish <-> Multiple Go backend servers LSWS - Enterprise 5.0.18 (same issue with 5.0.17 as well) Enable Public Cache No Check Public Cache Yes Max Object Size 256K Cache Expire Time (seconds) 3600 Cache Stale Age...

Hi, We cannot use Litespeed alone to achieve what we need for a couple of reasons: 1) Varnish acts as a load balancer in front of the ultimate back end systems (an application written in go), distributing the load of some CPU intensive processes over many application instances and servers. 2)...

We have Litespeed proxying request to a backend server, which happens to be Varnish - we do this because Varnish does not have SSL support and the Varnish backend has a lot more memory in it than the webserver. Unfortunately we are not able to purge items from the Litespeed cache, no matter...

Appreciate that, but we're talking about access control here. The access control should take priority over anything else: 1) People expect it to be 2) If the access control denies you, there's no point doing any more processing AllowOverride All is set (it's using Apache config file), also...

Hi, We've come across a serious bug with the processing of access restrictions (allow/deny) and .htaccess. Here's the situation: / .htaccess has a Rewrite to push all .php requests through a centralised router for legacy reasons. RewriteCond %{REQUEST_URI} ^(.*)\.php$ RewriteRule (.*)...

Downloads still not working though :/

Coming back to this issue, this still happens. It's only with iOS 8 + Litespeed. Use Apache and it's fine. We're not the only users experiencing this. Other users have reported it appears to be down to the two settings for: Enable Compression Enable Dynamic Compression Disabling them fixes...

As I suspected. I'm pretty sure it's down to the relative sizes of the images being requested inside the single pipeline - the smaller images will arrive before the larger ones presumably, even if requested later in the pipeline. iOS doesn't appear to handle it well.

Hi, We have a busy site that uses Litespeed and with iOS clients it keeps hitting issues with HTTP pipeline support, where images end up swapped - presumably because they have been sent out of the order in which they were requested. Most browsers seem to cope, but iOS devices seem particularly...