Painful browsing the forum, like wading through treacle

[#1]

Firstly I'm not disagreeing with you either, just brainstorming possibilities.

With your hypothetical case I have two thoughts -

If there was a single server doing all the work and for the sake of argument it had run out of RAM, and therefore was using a swap file to extend the memory available in order to accommodate the excessive load.

Hypothetically some user sessions might be in RAM and appear to perform normally, while other user's sessions may be operating entirely in SWAP and experiencing considerable lag.

Therefore the user experience could be massively different between simultaneous users.

If there was more than one server operating behind a load balancer, if one server was poorly configured but the others were somewhat better.  Then is is feasible that the user experience would be down to the luck of the draw, depending on which server the load balancer chose to allocate.  Extending this idea, the load balancer might be misconfigured and could hypothetically be providing poor balancing.

Unfortunately these are just hypotheses and we can only guess at solutions while Plusnet and Lithium continue to not actually do anything - after nearly THREE YEARS of complaints.

Plusnet need to allocate a competent engineer to examine both the Plusnet and Lithium sides of this problem, and if necessary find another forum vendor if no progress can be made.

[#2]

Taking your first single server hypothesis the lag would only be effective while the session details were fetched from the swap in to ram pushing something else out as a consequence. But again we can only assume that RAM isn’t the issue as any webmaster worth their salt would see this.

For the second one would assume, if not hope, that the machines used within a cluster would be hardware and software clones but that may not always be the case as older hardware is repurposed.

Indeed it is all just speculation as we’re not privy to the technical details this is all we can do. But it does surprise me that there appears to have been no obvious feedback from Lithium as I assume it would have been mentioned.

However, as @JonoH mentions in post 107 he says that Lithium are having trouble replicating the issue but then mentions the fact that the task has been assigned to ‘senior support’ staff, but why would you do that on something they say they can’t replicate? Or is it corporate speak for assignment to the bit bucket!

[#3]

Community has been offline for a few minutes. It now seems quicker. I wonder...

[#4]

I wonder indeed... 

[#5]

---

@Strat wrote:

Community has been offline for a few minutes. It now seems quicker. I wonder...

---

I agree, it does seem much improved, but still a little wait (for me) when clicking a board header.

[#6]

@JonoH

Your wonder rather implies that PlusNet's provider takes PlusNet's  service down for maintenance during core service hours … without checking it is OK to do so first?

Has Plusnet been advised of the changes applied … or is this a kick the tyres type of change?

PS: I can still see the paint drying!

[#7]

We escalated this issue today via our relationship manager, as it had been a week since our last update and the issue didn't seem to be moving forward. This issue was then almost immediately picked up by Lithium. When it went down we did wonder if it was anything to do with the investigation into the issue, they may be (and probably are) unrelated. 

The forum is lightening fast for me since the outage though.

[#8]

@JonoH,

I hope the Lithium understand what the restart clears out so that they can sort out what resources are being lost.

pingdom.com tests do report much better - sub 2 second page loads on the boards and the long wait times have been diminished, however there are a number of long DNS queries reported 150+ms and there is more than one of them!

Posting (especially amending) still seems very clunky though!

Really glad it is super fast for you!

[#9]

Definitely an improvement in reading posts.

Fingers crossed that posting this doesn't take long.

EDIT: Posting was fine.  

[#10]

Haven't we been here before?
(Temporarily suspend forum, reboot servers, hope that removes a few Klingons and thereby gives you long enough to proceed to nearest bar before 'the fan' gets switched back on. 🤔)

[#11]

Lightening fast right now!!

[#12]

Everything is much quicker for me, too, but I still see a noticeable delay on clicking individual board headers of 1-2 seconds.

Posting replies also has a little lag.

[#13]

I am keen to know what was done and what prompted the unscheduled restart.

[#14]

pingdom reports the same page to be 3% smaller load and 12 fewer HTML requests.

[#15]

Lithium has increased the memory available to the forum and things seem much better to us internally. How is it for everyone else?