Yes, we are aware; comment counts do not appear to be updating when viewing the main page.
We apparently have a resource availability issue. As far as I can tell, the counts are being updated correctly. (Yay!)
The problem seems to exist in a layer in which I lack expertise. The assistance of site admins has been requested.
UPDATE 20:23 23-02-2022
The problem has now been resolved. Many thanks to mechanicjay for fixing it again. As of yet we do not know the cause of the local mysql query node failure on one of our servers. However, as some of you have already deduced, there are many background tasks that are managed by the scheduler including updating comment counts, issuing moderation points, updating the journal index on the front page etc. The scheduler couldn't access the database and therefore couldn't do any of the tasks that we expect it to do.
In addition to expressing our gratitude to mechanicjay I would also like to say thanks to all in our community for your patience and good-humour in this rather protracted fix.
I claim this zeroth post for Natalie Portman and hot grits everywhere!
I claim *this* zeroth post. The next person can do the same...until it gets fixed!
Me too? Did Joe sniff you too?
I claim post ∞
Mine is way cooler.
I claim post 0oK. That's cooler.
I claim ∞+1 . . . . amateur
I'd just like to point out that no-one cares. There are zero comments on the story about comment counts.
It seems like the function that does that would
have been very stable. It's been working for years.
There can be only 0|1 explanation. It has something
to do with Soylent switching to secret new quantum
computing services. Henceforth, you shall not know
the number of comments until you observe it.
I remember that the very same problem already happened once in the past.
It has been happening about once every few weeks lately...
I wish I had time to grok the system well enough to have enough understanding to provide any sort of suggestion towards a proper, permanent fix...
We apologize again for the fault in the counts. Those responsible for sacking the people who have just been sacked have been sacked.
Døes that mean we'll get the rest øf the cømment cøunts in a cømpletely different style?
Sure, though the department responsible for switching that for you has been sacked. You have to do it yourself now. The steps are:
Asking for my friend Hariprasad in tech support.
New journal entries do not appear on the front page.
Perhaps this is a feature and not a bug.
Whatever process manages periodic functions on the site have fallen over.
Main page slashboxes like Journals are not updating, comment counts are not updating, peoples' accounts were not awarded moderator points at 00:10 UTC and probably a bunch of other things, both apparent and behind the scenes, are obviously not being run...
Unfortunately I understand very little of what goes on amongst the inner workings of this rehash-master directory I unpacked here as a learning exercise some time ago, it seems to be mostly a large jumble of random blib-blobs coded by rabid ferrets or something as far as I can yet tell. 😕
Some day perhaps I will have enough time to spend to decipher what goes on under the hood...
Perhaps it is the function that causes Varnish to update its cached version of the main page. Story pages seem to be updating since we're seeing new comments.
It seems that it must be more than just one process not being run, since things like the daily mod points system never ran either.
As an aside, comment counts have now updated, as has the main page Journals slashbox, but the mod points thingy hasn't run yet (or is perhaps processing now but not done yet) but presumably that might have to be started manually once the assigned run time has passed if the process normally in charge of running it was a state of borkedness?
/me knows not
Looking at this story [soylentnews.org] that claims 5 posts, it appears whatever process that stopped working it happened between 10:22 UTC and 10:37 UTC (assuming I can count to 5 correctly). That'd work out well for a process that kicks off on the half hour getting hosed and not completing, thus preventing future executions of that process from launching.
It is the slashd daemon. Instead of using a cron daemon to run scripts that do various pieces of site maintenance, they wrote their own replacement. That was not necessarily a problem on its own, but the system it is set up on doesn't have any sort of watchdog or service management. So when it goes under, there is no automatic maintenance or failure notifications if they don't notice the problems on their own.
I'm surprised no one replied to this yet. Seems like a big deal.
I also didn't receive mod points this morning and the messages I would usually receive about yesterday's posts did not arrive.
As I recall, when this occurred previously and the script that gives mod points didn't run, it wasn't manually run at a later time in the day. Instead, nobody got mod points until the next cycle at 00:10 UTC. If that's how things are done now, it'll probably be about seven hours until you receive a new supply of mod points. I have no inside information, but that's what I recall happening previously when this was an issue.
I didn't receive any either.
The back pressure of mod points may cause damage to the plumbing.
Ideally a server should have periodic replacement of vacuum tubes, belts, worn parts and lubrication of gears and bearings.