Ghost / MEW front page going offline for an hour

I will leave that for the brainiacs :smiley:

1 Like

I’m with @Wyvern here.

Why is snapd, mysqld, and kswap chowing so many resources?! :’(

is all just weird mouth noises to me.

Will copy your post over to the Tech Support topic in case the clever ones don’t think to look in here.


Maybe check logs? Not sure which ubuntu version you’re running, but I assume systemd is available.

journalctl --system -f -u snapd

journalctl --system -f -u mysqld


I missed this, but you can set snap refresh timer. View with snap refresh --time
Change it to maybe daily at 4am? Only person up, seemingly is @GregRedd :joy:

sudo snap set system refresh.timer=4:00

Maybe confirm my commands, I don’t use ubuntu. Last actual install is 18.04


Am amused by how perfectly positioned the MEW rhino is on the 2560 monitor :smiley:


Not sure if my phone fault or a MEW problem. My keyboard now blocks the typing block when replaying from my phone. It used to slide above the keyboard.

Will be playing around and see if it is just me…


Nevermind… Works on Chrome but not Edge…


Howdy folks. I see the new upgraded Ghost is still giving some trouble.

Now that the Ryanair chaos from last week seems to be over, I’ll spend some time to see if I can make it more reliable.

Thanks for the tips, @aldyr.

Curiously, the refresh.timer property isn’t currently set, so I wonder what Ubuntu 20.04’s default is and if maybe Digital Ocean doesn’t have a different way of handling that in its images?

I’ll muck about with it today and see what I can see.

My first port of call will probably upgrading the instance from their “standard” plan to a “premium” Intel or AMD plan. Now I just have to figure out whether we should go for a newer-gen Intel Xeon or AMD EPYC config.

1 Like

Looks like AMD is the winner:

But I’ll wait til tonight to do the upgrade so y’all can write your articles in peace.


Still getting warnings from Google about downtime when the auto update runs, even after the upgrade to newer AMD hardware.

One could just have the auto updater run once or twice a week, but I think it’s better to just upgrade the instance at this point. Google doesn’t like downtime — your search result rankings get downgraded for it.

I’ll upgrade the instance tonight after everyone’s published their articles and stuff. It’ll be more expensive, but it’s the best solution as far as I can tell.


I guess it’s too late to rollback to the earlier, less resource intensive, less expensive version?


I still have the backups so it’s probably possible. But we can’t run unsupported versions of the software — it’s just too risky.

It’s the difference between around R80 per month and R160 per month, so it won’t plunge me into poverty or anything.

One last thing I can try is see if there’s some way to tune MySQL to use less RAM. Maybe the default config in Digital Ocean isn’t good.


You must let us know if we can help with the payments