5 second lag when same image called multiple times in a short period of time

[wanah]

Hello,

I'm trying to find the setting that would cause a 5 minute lag when the same image is called about 10 times in a row from within a PHP script that's not CPU, memory or disk limited.

One of our clients scripts does this and recently he's experiences lags over 5 seconds because of this.

I belive the server's IP is whitelisted and I know litespeed webserver is far from using it's current allowed 4 cores.

I've been through all the settings and things like make connections per client is set to 200 so this shouldn't be the limit that is hit. Here are my settings :

Max connections : 5000
Max keep alive requests : 10000
Max static requests per second (per client) : 200
Connection soft limit (per ip) : 80
Connection hard limit (per ip) : 100
Block Bad Request : No
Enable Request Filtering : No

What am I missing ?

 

[NiteWave]

5 second or 5 minutes ?

>the same image is called about 10 times in a row from within a PHP script
can you show the example php script ? let's see if can reproduce at our lab.

 

[wanah]

Hello,

I haven't tried out of context.

I'll try and reproduce the issue on a test account and let you know.

It's 5 seconds not 5 minutes.

It does the first 15 images in milliseconds then waits for 5 seconds before doing the 16th image then does the next four images in milliseconds.

It doesn't happen every time (on average once evey 5 times) and never on the same call (because it's 20 times on the same small image hosted on the same account and it seems to block for 5 seconds when it hits a limit).

 

[NiteWave]

it looks an interesting issue.

you can do the tests on the server directly to compare the result.

just use ab (apache benchmark). for exmaple,
#ab -n 20 -c 1 website.com/img1.jpg