Did we blow up Strava already?

[jrenaut]

@hozn 151334 wrote:

There is definitely a freezing saddles software issue. I have been ignoring emails all day. Apparently the (MySQL) database crashed and one of the tables needs to be repaired. It is unfortunate having to use such a horrible database

Anyway, I will get that fixed tonight.

Perhaps it’s time to take up a donation to get better hosting? I’m really happy with Heroku, and we could probably do the free tier of Postgres along with a $7/month app dyno (and maybe even free SSL). Seems like we could pass the hat for $84/year.

[peterw_diy]

@hozn 151334 wrote:

There is definitely a freezing saddles software issue. I have been ignoring emails all day. Apparently the (MySQL) database crashed and one of the tables needs to be repaired. It is unfortunate having to use such a horrible database

I know you despise MySQL but the Strava app and widget misbehaving today suggests to me that the problem likely stems from them.

[jrenaut]

MySQL is always to blame. Unless you’re using server-side Javascript and then you’re to blame.

[hozn]

@peterw_diy 151337 wrote:

I know you despise MySQL but the Strava app and widget misbehaving today suggests to me that the problem likely stems from them, not BAFS’ hosting. Maybe there’s a related problem like your batch code not elegantly handling problems with Strava’s services, but I’m skeptical of the idea that MySQL is to blame here.

Well, Strava uses MySQL too, so I’m not sure we can leave MySQL out of this either way

But, yes, it sounds like a few things might be happening here. Unfortunately I have to use a mixture of ACID InnoDB table and the apparently-easily-corruptible MyISAM tables (for geospatial indexes), so I’m not exactly sure how transactions span operations in the different table types, but I know my ride_geo table has been in a needs-repair state for a significant part of the day.

(It is now repaired, so hopefully any remaining issues lie squarely in Strava’s court.)

[Judd]

@hozn 151339 wrote:

Well, Strava uses MySQL too, so I’m not sure we can leave MySQL out of this either way

But, yes, it sounds like a few things might be happening here. Unfortunately I have to use a mixture of ACID InnoDB table and the apparently-easily-corruptible MyISAM tables (for geospatial indexes), so I’m not exactly sure how transactions span operations in the different table types, but I know my ride_geo table has been in a needs-repair state for a significant part of the day.

(It is now repaired, so hopefully any remaining issues lie squarely in Strava’s court.)

Looks like my rides from today synced on the leaderboard. This evening’s Hains Point laps haven’t updated yet. I really appreciate all of the work that you do to make the game possible Hozn.

[hozn]

@jrenaut 151338 wrote:

MySQL is always to blame. Unless you’re using server-side Javascript and then you’re to blame.

I couldn’t agree more. Any language that needs a library to check whether a value is an array (https://www.npmjs.com/package/is-array) really should probably not be used for anything.

It’s one thing if it’s the only option, because that’s what browsers support, but when people knowingly choose to use that horrendous language on the server-side !?! I have been writing a lot of ESNext script (ECMAScript 2016+) for a current project and while that is a whole lot better than old-school JS, transpiling introduces a new layer of problems.

[hozn]

@Judd 151340 wrote:

Looks like my rides from today synced on the leaderboard. This evening’s Hains Point laps haven’t updated yet. I really appreciate all of the work that you do to make the game possible Hozn.

Glad it is working! — And, no problem — despite my whining, I hope it’s obvious I do enjoy this stuff. I wish I had more time to contribute to it. This year is looking particularly difficult given workload [at work].

[ewilliams0305]

@Vicegrip 151297 wrote:

Can we ask Eric B to do something about his elevation error issue?

Don’t worry, as soon as I stop bullshitting around I’ll catch him with an actual bicycle, with actual hills, with actual pedals[emoji1338]

Regarding the fix, it’s really simple, log on via a web browser and click the ? Next to elevation then click resync elevation. It really does spoil the chart for those of us that actually enjoy climbing:(

Sent from my iPhone using Tapatalk

[peterw_diy]

@hozn 151341 wrote:

I couldn’t agree more. Any language that needs a library to check whether a value is an array (https://www.npmjs.com/package/is-array) really should probably not be used for anything.

The existence of a npm package/project is not evidence of its necessity. Don’t forget the left-pad debacle! http://www.theregister.co.uk/2016/03/23/npm_left_pad_chaos/

[jrenaut]

@hozn 151341 wrote:

I couldn’t agree more. Any language that needs a library to check whether a value is an array (https://www.npmjs.com/package/is-array) really should probably not be used for anything.

It’s one thing if it’s the only option, because that’s what browsers support, but when people knowingly choose to use that horrendous language on the server-side !?! I have been writing a lot of ESNext script (ECMAScript 2016+) for a current project and while that is a whole lot better than old-school JS, transpiling introduces a new layer of problems.

var whee = Array(‘bicycles’);
var wtf = Array(2);

Now whee is an array with one element – the string ‘bicycles’. And wtf is an array with two null elements. THIS IS NOT OKAY.

I mean, it’s cool that you can assign anything to a variable. I’m pretty sure var mr_peepers = literally a real monkey that you stole from the zoo; is ‘valid Javascript’ and you can pet Mr Peepers and he’s adorable and all that and he inherits all the right properties. But it’s like if you’re writing a novel except every few pages the letter “a” is treated as an imaginary number. I’m sure your novel is JUST GREAT but I don’t want to read it.

[Judd]

@ewilliams0305 151343 wrote:

Don’t worry, as soon as I stop bullshitting around I’ll catch him with an actual bicycle, with actual hills, with actual pedals[emoji1338]

Regarding the fix, it’s really simple, log on via a web browser and click the ? Next to elevation then click resync elevation. It really does spoil the chart for those of us that actually enjoy climbing:(

Sent from my iPhone using Tapatalk

It also spoils the chart for us that don’t enjoy climbing but enjoy looking at charts of how much non-enjoyable climbing we did.

[hozn]

@peterw_diy 151344 wrote:

The existence of a npm package/project is not evidence of its necessity. Don’t forget the left-pad debacle! http://www.theregister.co.uk/2016/03/23/npm_left_pad_chaos/

It is true that in newer JS you don’t need that package, but his was traditionally a not entirely obvious problem; we used to ask this as an interview question to see if candidates really had js experience. (Everyone puts it on their resume.)

That and what is the value of x and y.
var x = “1” + 2; // “12” of course!
var y = 1 + “2” * 3; // 7 of course!

[Steve O]

@jrenaut 151336 wrote:

Perhaps it’s time to take up a donation to get better hosting? I’m really happy with Heroku, and we could probably do the free tier of Postgres along with a $7/month app dyno (and maybe even free SSL). Seems like we could pass the hat for $84/year.

If this is really a thing, put me in for $5.

[Judd]

@Vicegrip 151297 wrote:

My am ride not showing up ether. Don’t seem to recall Stava being at all laggy in past years.

In other news. Major Tom is back again this year on the dist/elev/ave speed chart. Can we ask Eric B to do something about his elevation error issue?

Just posted on one of his rides letting him know that he’s breaking one of the charts and asking him if he could correct it.

[hozn]

@Judd 151355 wrote:

Just posted on one of his rides letting him know that he’s breaking one of the charts and asking him if he could correct it.

If this keeps happening we can drop out-of-tolerance elevation numbers unless it has been corrected (I believe that is a flag on the activity).

[Author]

Posts