After several days of seemingly endless outages and timeouts, We Hunted the Mammoth is on a new server and … working ok so far (crossed fingers). There’s one more little change the WordPress people need to make; hopefully that won’t disrupt the site for too long.
In any case, I’ll resume posting shortly. Welcome back, everyone!
Follow me on Mastodon.
Send tips to dfutrelle at gmail dot com.
We Hunted the Mammoth relies on support from you, its readers, to survive. So please donate here if you can, or at David-Futrelle-1 on Venmo.
Like the election. Am I right folks?
But seriously, congratulations. That must be such a relief.
It’s kind of not? I keep getting “500 Internal Server Error”.
I was getting 500 errors for days, but as of this posting, it’s working.
Avoiding discussion of the fascist manosphere’s latest bloviations was probably good for me anyway.
I had a hunch there was a technical snafu somewhere down the line. Thank goodness it’s been straightened out.
On an unrelated note, former Japanese PM Shinzo Abe was shot earlier tonight and the prospects for his survival are not looking good.
I am still having issues where I get to the site but hitting any particular blog entry is a crap shoot. Sometimes it takes a while to come up, sometimes it gives me an internal server error. I dunno, for what it’s worth, I’m not that worried about it, just mentioning it.
@ Chris Oakley
Yes, and the gun people over here in the US are all “PROof Gun control doesnt work!!!”
Looks like you can count the number of gun deaths in Japan on your fingers. Don’t see how they come to that conclusion. We Americans probably had more shootings this morning in Chicago than that.
Can’t wait to see Ms. Jewish Space Lazers weigh in on it…
@Chris Oakley:
And Abe Shinzō was declared dead as of 5:03 PM Japan Standard Time/4:03 AM US Eastern Time: https://www.cnn.com/asia/live-news/shinzo-abe-japan-pm-collapses-nara-07-08-22-intl-hnk/index.html
Yay!
@Victorious Parasol
Interesting timing on that post XD
@An Impish Pepper
Heh. Well, I gotta confess, I have no special sentimentality to former PM Abe, though I did know a couple of Japanese students at uni who were there to get an education and then go back to Japan, and I hope they are well. We’ve got a surplus of global instability right now, and I’d like to be bored for the rest of this year, please and thank you.
Former PM Abe’s assassination will probably have an influence on upcoming elections.
Still loading more slowly than any other site I visit.
A thought regarding Hugh Grant’s music request (and Steve Bray’s gleeful compliance) for Boris Johnson’s outro: a case could be made for the Benny Hill theme as an aspirational choice.
Benny at least knew how to multitask despite perverse and antiquated technology and how to take decisive and resourceful action in the face of colliding crises:
http://m.youtube.com/watch?v=MItPs6iAGiU
Thanks for the feedback. I guess the way to put it is that the site is half-fixed; they got it on a new server and added RAM but it still has problems with a balky database (thousands of posts and hundreds of thousands of comments). That’s why it’s slow. Hopefully I can get someone to dive in there v. soon and streamline the thing, or whatever you do to fix a database without deleting massive amounts of old content here.
… or I could just change the site’s tag line to “slowest site on the internet.”
@ david
… or I could just change the site’s tag line to “slowest site on the internet.”
As Ridley Scott said “What’s wrong with a bit of f***ing build up.”
The website is still slow and unreliable on my end.
I still get error when coming onto the site.
Only works on my phone
Works, but sooooo slllloooooowwwwwwww …
Your new server seems to be a downgrade. Like going from a modern PC to a 90s-era 80486 with emulated floating-point. Ugh!
Let me guess: budget cuts?
OT:
1. Does anyone else find things like this annoying?
You come across an interesting looking blog link like this:
https://freethoughtblogs.com/stderr/2022/07/09/an-explanation-for-inflation/
You click it. And there’s no meat there. The text on that page does not contain anything remotely resembling an explanation for inflation, or whatever the topic was that interested you. Instead, all it says is there’s this link to this audio file, or Youtube video, or some such that in turn purportedly contains said explanation. So, maybe the thing actually exists but you won’t be able to read it, stop to think here and there, etc., and you certainly will not be able to control-F for things in it; it will automatically go at its own predetermined pace, regardless of your wishes, and it might cost you an arm and a leg in mobile data while it’s doing so.
If I were looking for Youtube videos I’d search Youtube. If I were looking for audio files I’d probably use Soulseek or something like that. If I’m looking at blog links I was searching for text, something I can read and that is lightweight bandwidth-wise. Pah!
2. Does anyone know of a superior alternative to Allegra? Not only is it expensive at ~$1 a day, the packaging is really inefficient (huge amount of single-use plastic waste per gram of product) and it is proving to be just plain unreliable. It has an unpleasant behavior where taking one appears to instantly cancel any remaining effects of the previous dose, but the new one takes some tens of minutes to fully kick in, so there’s an unavoidable 20 minute or so window of being unprotected each day; so, every day after I get up and take that day’s pill, I get to spend most of the next 20 minutes sniffling and being distracted by itchy tickling in my nose before the new dose kicks in. But making matters worse are the occasional duds. Two days ago, for example, it was nonstop sniffle, sniffle, sneeze, sneeze waiting for it to take effect, which it never did, or at least not until six or seven hours later. And I don’t know then if it would be safe to take another. Maybe the dud is missing only one of two separate pharmacologically active ingredients, and taking another will result in overdosing on the second? Plus, as I said, it’s a full extra dollar if I do that.
3. Is DeviantArt still “where it’s at” for an online art community or has it become a flaming garbage pile?
Folks, I appreciate your patience as I try to get the site fully fixed. Right now, as I’ve said before, it seems as though the root of the problem is a database problem, and I’ve got someone lined up to take a look at that and any other issues that might be causing the slowness.
FWIW, the site does seem to work a little faster on tablets and phones. At least that’s been my experience.
There are no “budget cuts,” and the new server is not secretly an old server.
@Surplus: Allegra is an antihistamine. So you could probably replace it with any other antihistamine. I personally like the generic version of Zyrtec that I can find at walgreens (called Wal-Zyr) – though other generics likely exist in other places.
Whether an antihistamine works well for you is a bit of trial and error. I know my mom reacts best to a different one than me.
It does take a few minutes, at least, for an oral dose to distribute into your blood stream. So it is not actually possible to have an immediate reaction. And on some days when there is more allergen/your body is producing more histamine there may be a weaker effect.
I know with Zyrtec a double dose may be safe (doubling was recommended to me by my doctor but may not be safe for everyone). Most OTC drugs have a pretty hefty safety margin built in. If you are a smaller adult it may be less safe for you to double the dose.
As a fan of delayed gratification, I don’t mind at all that the site won’t be rushed into loading.
It makes me shiver with antici…
If the problem is the database then would it maybe be helpful to archive the old threads on another site? Just asking, don’t know much about conputers.
Archiving old stuff hurts searchability (now 2 places to search and not 1) and shouldn’t have any noticeable benefit. Databases are generally built on B-trees, whose performance worsens only logarithmically with size.
Further evidence against size being the cause is that the appearance of problems occurred in a step-function fashion some time on July 2, rather than developing gradually over a very long time, the way a logarithmic drop in speed would manifest.
Someone did something on July 2, perhaps a subtle mistake like adding a feature that performs a query that keys on a column that the database wasn’t told to index (i.e., build and then maintain a quickly searchable B-tree for), resulting in a linear search whenever that query is run. The mistake might be either keying on that column at all, or else not building an index for that column now that you’re going to be keying on it. There are other things that could be problems (e.g. outer joins with a lot of matches can produce quadratic memory demand) but not indexing on a column that’s newly being used as a key is perhaps the most common error that would lead to a drop in performance. Hitting a pathological edge case in the vendor’s specific implementation is even a possibility, e.g. the way a stripey pattern of alternating oldest/newest can make quicksort-type algorithms take quadratic time, or data with a particular pattern might cause a poorly chosen hash function to put most of it in only a few buckets. The vendor would have to fix it in this case (e.g. replacing quicksort with mergesort or changing a hash function) or the user (here, David) would have to return to avoiding that edge case. Not knowing the details of what was being done on July 2 that precipitated all of this, though, all I can do is speculate and suggest some things to be on the lookout for. It helps that one need only go over the things touched by items in the July 2 changelog instead of the whole system looking for these sorts of things; the culprit change is somewhere in there, since that’s the day the problem developed. Well, this is true unless the vendor has an outright bug that caused what should have been an innocuous action to trigger the database to shit the bed in some manner … a log of all database transactions from July 2 will still contain the triggering action, but it won’t necessarily look significant, let alone be erroneous in character. Instead, the software’s response to it was erroneous in this case. Finally, if the cause of the server migration was a failing disk drive, specifically, then something (e.g. a B-tree index) could have been corrupted and might still be corrupt after copying the affected files to the new drive in the new machine. There should be a way to make a database rebuild its indices from the raw contents of the database; this operation will likely take a while (linear or even n log n in the database’s size) but it should fix any corrupt index by replacing it with a freshly generated version. Profiling to find which specific query is being abnormally slow might narrow down the problem to queries that key on a single particular column in a single particular table, in which case you know it’s probably a particular index that ought to be rebuilt and needn’t rebuild all the others.
If there’s a full, from-inception transaction log of the database, then the nuclear option (likely very slow) is to rebuild the whole database from scratch by playing back that log into a new greenfield database, then running all the new-since-this-started transactions on the old database onto the new one, and repeating until fully caught up, then swapping in the new database as the “live” version. That should eliminate any corruption of the database’s structure. It won’t help if the pattern of data is hitting a slow edge case in the implementation, though (but if it doesn’t fix it, then that’s almost certainly what’s happening, and that problem pattern of data must only started to exist on July 2, for whatever reason, making this seem less likely).
I’ve been reading the site on my laptop tonight and experienced some slowness and inability to load (primarily on the Amber Heard post), but it now seems to be resolved pretty easily by disconnecting and reconnecting to my WiFi network.