Improve readability and completedness for Redis

[martin-rueegg]

• give more complete examples for Redis config
• use separate sections for the different connection types (TCP, TLS, socket)

Signed-off-by: Martin Rüegg martin.rueegg@metaworx.ch

[blizzz]

Great improvement, few nitpicks :)

[martin-rueegg]

Thanks @blizzz for taking the time for reviewing this.

[martin-rueegg]

@nickvergessen, @blizzz, @CarlSchwan are you happy with the changes I made following your request?

[nickvergessen]

I'm fine but I have little to no background to verify the content of your changes

[martin-rueegg]

@PVince81 @skjnldsv , any chance to have a look at this, please?

[szaimen]

@blizzz can you approve? :)

[martin-rueegg]

@PVince81 @szaimen @skjnldsv

All points raised by @blizzz regard information/sections, that have already been there before. I am not able to technically review the content of the entire document. But if additional changes are required then I suggest to proceed as follows:

• If the change regards the formatting or layout or structure of the article, I'm happy to adjust those.
• If the change regards the technical content, then a new PR/issue should be opened to address those.

Otherwise, we keep on the improvements in this PR from being published, and potentially helping users/admins. Just because the document should have been overhauled/redacted technically already.

What do you guys say?

Addressed

[martin-rueegg]

Thanks, @skjnldsv.

[skjnldsv]

Thanks for your patience @martin-rueegg 🤗

[ChristophWurst]

I found

Using Redis for local cache on a multi-server setup can cause issues. Also, even on a single-server setup, APCu (see section above) should be faster.

for https://docs.nextcloud.com/server/latest/admin_manual/configuration_server/caching_configuration.html#redis-configuration-in-nextcloud-config-php but not https://docs.nextcloud.com/server/25/admin_manual/configuration_server/caching_configuration.html. Does that mean with Nextcloud >= using Redis for local cache is not problematic? Otherwise I would suggest to backport the refinements of our recommendation.

[ChristophWurst]

My remark is specifically about b502cc7. We can also only backport this one commit and leave the restructuring for 26+.

[martin-rueegg]

My remark is specifically about b502cc7. We can also only backport this one commit and leave the restructuring for 26+.

It's funny, because when I created this PR back in July last year, versions 22-24 where actively released. Don't think any of the examples I've added have been unavailable at that point. So why would it not be back ported to at least v24 or even before?

The note you mentioned was added due to @blizzz 's comment.