slowwww

Founder
Joined 2000
Paid Member
Our tests show that it is related to networked filesystem read speeds, and moving to a new DC will certainly prove or disprove that theory. As well as being something that was already desired and being planned for on the back-burner (moving to a cluster with a higher level of monitoring and 24x7 problem resolution) - we have just brought forward that migration as right now would be a great time to kill two birds with one stone.
 
  • Like
Reactions: 5 users
I thought it was my poor internet connection
 

Attachments

  • GTmetrix.jpg
    GTmetrix.jpg
    165.2 KB · Views: 54
Account Closed
Joined 2018
Well, it's time for me to un-shelf that integrated amp project that I've has sitting since 2019.
I'll use the "offline" time to cut and drill and prep the chassis which is the tough part.
The various modules have been tediously done and tested since 2020 - power amp/B+/B- supply, preamp/tone. RIAA pre, standby/on-off, etc, etc.

Good luck on the migration.
 
as the issue started when they had a connectivity issue on March 4th..

I noticed already on March 2nd something was wrong when I tried to upload several small sized images in my "Avatar issue" thread which would stall the upload several times and there was no other way than hit F5 as the page became unresponsive, but recently browsing around on DiyA have gotten much worse.

My thought was if there's some large scale cyber attack going again but can't find much news.
 
Our tests show that it is related to networked filesystem read speeds, and moving to a new DC will certainly prove or disprove that theory. As well as being something that was already desired and being planned for on the back-burner (moving to a cluster with a higher level of monitoring and 24x7 problem resolution) - we have just brought forward that migration as right now would be a great time to kill two birds with one stone.
Personally, I would check PHP`s session files (mostly TMP dir) and clean them. There could be a problem with inodes. If every unique visit creates a session, that would be a real issue on forum this big.
 
Administrator
Joined 2004
Paid Member
LOL GTmetrix itself is slow and stalling out now.
Obviously we have problems here that necessitate a data center move, but I’m not sure the GT is the best tool. It gives a grade D to a website that is always super fast loading for me.
 
Neurochrome.com
Joined 2009
Paid Member
For those using GTmetrix: Beware that it defaults to testing from Vancouver, BC, Canada 🇨🇦. That's a lovely place, especially in the summer, but not so useful if you want to test the performance from other locations.
Screenshot 2023-03-08 at 10.31.49.png


If you want to test from other locations you need to create an account with them and change the testing options. Creating the account is free. Then you can do stuff like this:
Screenshot 2023-03-08 at 10.48.14.png

So pretty much universal suckage world-wide, which is congruent with what everybody here is saying.

And, of course, the point of the free account is that GTmetrix then gets to sell you a PRO version with even more options. I guess they have to eat periodically too...

I'm not associated with GTmetrix. I just use their (free) product and find it useful. I've used it in the past for debugging my website and also to hold web hosts accountable. It was one of the better web hosts that made me aware of the tool in the first place.

Many thanks to Jason et al. for your hard work keeping the forum running and fixing the various issues.

Tom
 
  • Like
Reactions: 1 user
Seems to a problem on Gmail as well, Google's other sites like YouTube are also noticeably slow, I am not seeing movies in 720p, even when I ask for it, stays at 480p resolution. 360p at times...

There could be many reasons, slow fiber is only one of them.
We are two miles away from a small town and at the end of the cable for TV, phone and internet. There are 5 houses past us, and after that there is no wired or fiber internet at all. Service interruptions and outages are common. diyAudio has been sluggish and unusable at times for the last few days, but all of the Google stuff is fine. I can watch a 4K / 60 video full screen without any dropouts or stuttering just about any time except for weekdays from about 6 to 8 PM, which is normal here.

The diyAudio issues are now well known, but your Google issues are likely due to something else, but your server for Google services is likely not the same as mine.

I just watched a nearly 3 minute 4K video and came back to this page to see that it had not loaded yet.
 
I use a 4G cell phone as the modem, via tethered USB.
I will have to see if I can tweak the DNS settings to improve the speed.

Fiber optic lines have been installed in my housing development, I will have to ask the neighbors about their experience.
Most of us use satellite TV receivers, as our country is seeing a lot of construction, buried cables getting cut is common, and also an issue during the rainy season here.
Cell phones are more than 90% of connections, land lines are expensive and are up only about 4 months annually (from the state owned service provider).
Many people use smart phones in place of computers and TV...people watch movies on them.

The fiber optic package includes cable, internet, and an end unit that works as w ifi hot spot as well. A wired telephone can also be used.
It is about the same monthly fee as its wireless competitors.
 
Good Luck...
Do they have dual redundant servers for safety?

I am not sure what Jason is going to be using.

One aspect of my work is a project work associated with power monitoring software and hardware installation and commissioning in large data centres in Sydney (like Amazon, Equinix, NEXTDC, Global Switch...). I can tell you that these data centres' obligation is to provide an uninterrupted supply (mains voltage, in case of wide power outages.... they use large diesel generators usually located on the top floor of the data centre building; the generators can run indefinitely if required - therefore supplying the computers installed in PODs with mains power -> to preserve data).

  • The customer's obligation is to take care of the computer system (installed in POD); they can choose what they want to install and how to use the hardware, i.e. the customer is in charge of their computer system data availability (redundancy). The customer can install whatever software/store whatever data they want.
  • The data centre guarantees stable and uninterrupted mains power - and nothing else.

....... I thought I might kill some time while waiting for the speed to recover, by providing the above info... many are most likely aware of it...
 
Last edited:
We all work and play within the office at times, so we assume all is known.

I thought it would be server farm, so data storage and transfer would be through a common facility, with the possibly dedicated servers being provided by the facility, owned and maintained by them.
You are describing a rented building with backup power supply and possibly climate control.

I did make a website for my products, I paid for the domain name to one company, and another charged me a modest amount to host the site, and gave me a software called Site Studio to build it.
That was long back, I stopped renewing it because I did not get many visitors.

Another issue is that the fiber optic undersea cables, and their interconnects to land based networks, particularly in India, are also going through an evolution, so sometimes they run slow.
And the explosion in data is huge, many people here use up 3GB daily on cell phones, the data handling has to meet the demand.

Now that is common for all sites, so for reasons best known to the admins, this particular site has been slow lately, compared to other sites.

By dual redundancy, I mean severs in two separate locations, so a calamity does not put you out of service for ever.
It has happened, a data center caught fire in Europe, some members here had their websites hosted from that center.

The Indian Railways has a company called IRCTC, which handles catering and ticket booking.
Their ticket capacity is being expanded to 225,000 transactions a minute, and their base kitchen near Delhi makes 150,000 meals daily.
They have servers in different cities, in parallel, and in physically safe locations, one is near a military facility, another in an abandoned mine, so the story goes.

So, we differ, politely, in our concept of dual redundant.

It should be possible to use servers in different continents to handle the load from nearby locations, and share data globally.
That is more your area of expertise.
I am more comfortable with physical objects like cars and plastic molding.