Jump to content

Recommended Posts

Posted

They are not.

In the case of sessions - those are always temporary even when using the database, so them being "lost" is not an issue. If the user associated with the session has opted to have their login remembered, then it will just create a new one. Otherwise, they would just need to login again.

For views - a task runs through these and updates content items view counts accordingly. If these are lost for some reason before they are applied by the task, then they will not be processed.

Posted
8 minutes ago, Ryan Ashbrook said:

They are not.

In the case of sessions - those are always temporary even when using the database, so them being "lost" is not an issue. If the user associated with the session has opted to have their login remembered, then it will just create a new one. Otherwise, they would just need to login again.

For views - a task runs through these and updates content items view counts accordingly. If these are lost for some reason before they are applied by the task, then they will not be processed.

Cool ok.  Do you know how often that task runs so topic views on the forums reflect the most recent values?

Out of curiosity is this a big mysql load savings?  Thanks!

Posted

Every 5 minutes.

And yes, it can be beneficial particularly for sites with a lot of simultaneous traffic as it'll offload all session handling from the database. Combined with guest page caching, it's entirely possible to serve a page to guests entirely from Redis and never touch the database.

  • 1 month later...
Posted
On 12/22/2020 at 6:47 PM, Ryan Ashbrook said:

Every 5 minutes.

And yes, it can be beneficial particularly for sites with a lot of simultaneous traffic as it'll offload all session handling from the database. Combined with guest page caching, it's entirely possible to serve a page to guests entirely from Redis and never touch the database.

IPS is ultrafast for guests when using redis. A logged user browsing experiencie is clearly much slower.

Posted

We turned off "Redis to reduce MySQL overhead" in our testing 4.5 and before upgrading from 4.4. It was significantly slower when logged in. Maybe we have something not optimized correctly. We did not have that issue with memcached, which is what we were using in 4.4.

Posted
3 hours ago, Paul E. said:

We turned off "Redis to reduce MySQL overhead" in our testing 4.5 and before upgrading from 4.4. It was significantly slower when logged in. Maybe we have something not optimized correctly. We did not have that issue with memcached, which is what we were using in 4.4.

We have made some optimizations for 4.6, mainly disabling Redis data being encrypted by default.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...