Could the 10 day time out be extended on our own xf forum?

[Dean]

Any technical downsides if we wanted to make the current 10 day timeout much longer? 6 months perhaps?


Any large differences in performance or server load?

(on our own xf forum, not suggesting a change for xf.com)

 

[Mike]

Yes, it just stores more data and will make some of the automatic read checking behaviors slower. But it really depends on the volume of posts (well, threads) on your forum, as well as the number of active registered users.

 

[Dean]

Thanks!

Think we will start with a long timeout for usability, many (important) people only visit every few weeks. Then reduce it if necessary. I'd rather start long, and reduce if necessary, than the other way round.

 

[Boothby]

May I ask what 10 day timeout you are talking about?

 

[Dean]

May I ask what 10 day timeout you are talking about?

After that amount of timeout, currently 10 days on xf.com, posts are marked as read - even of they have not been read.

 

[Boothby]

Thank you for explaining.

 

[Shanj]

After that amount of timeout, currently 10 days on xf.com, posts are marked as read - even of they have not been read.

It also means that once posts are marked as Read, a user opening a New Thread (to that user) gets bumped to a post near the middle or end of the thread.
This is confusing as it's not intuitive to open a new thread and it doesn't open at the first post. It's OK if it's obviously an old post like 6 months ago maybe but otherwise it's very confusing.
So yes it's important that we c an set the timeout as a lengthy time period.

 

[Forsaken]

It also means that once posts are marked as Read, a user opening a New Thread (to that user) gets bumped to a post near the middle or end of the thread.
This is confusing as it's not intuitive to open a new thread and it doesn't open at the first post. It's OK if it's obviously an old post like 6 months ago maybe but otherwise it's very confusing.
So yes it's important that we c an set the timeout as a lengthy time period.

You already can do that from what Mike and Kier have said.

 

[Kier]

On a very busy board with a high turnover of content, setting a high limit could result in very large quantities of data and a possible performance penalty - however on these kind of boards there tends to be less thread necromancy, so the problem is mitigated and a smaller timeout is less of an issue.

On a small to medium board that has plenty of disk space and doesn't tax its server, you could easily set a timeout of a few months without a problem.

 

[Floris]

On a very busy board with a high turnover of content, setting a high limit could result in very large quantities of data and a possible performance penalty - however on these kind of boards there tends to be less thread necromancy, so the problem is mitigated and a smaller timeout is less of an issue.

On a small to medium board that has plenty of disk space and doesn't tax its server, you could easily set a timeout of a few months without a problem.

Post cache and APC will be your friend - or in this case, won't that matter? (talking about 28 days on big-boards)

 

[Mike]

Post cache and APC will be your friend - or in this case, won't that matter? (talking about 28 days on big-boards)

Really has no relationship...

 

[Dean]

We are now using vb3.8, and have a 30 day timeout and have zero problems that I know of.

Testing how much load a longer timeout value would add seems extremely difficult to me, if for no other reason it would be a large inconvenience to people.


Could an xf forum work (ballpark) similar to vb3.8 with regards to the timeout vs load/speed, given the same conditions?

If you cannot say, I completely understand.

 

[Kier]

Should work identically, the concept of the read marking system is fundamentally the same.

 

[Freddie]

... thread necromancy ...

XenForo practices the black arts