SN slowness apology \ stacker news ~meta

pull down to refresh

3075 sats \ 37 comments \ @k00b 29 Nov meta

During a database backup last night, the filesystem where we store temporary (non-critical) database files got borked somehow. For some reason, the database's cope, likely recreating the temporary files over and over again without anywhere to store and reuse them, caused it to get really slow. It took us awhile to figure out that was the cause and it was fixed by literally turning it on and off again.

Thank you to everyone that reported it, and sorry for the downtime and the verbose error messages. (It's not wholly intentional to show you the sausage making but it helps us diagnose issues easier until we create some kind of error digest system.)

view all related items

175 sats \ 1 reply \ @DesertDave 29 Nov

It's so awesome that we get to hear about this directly from the guy who maintains it. You guys rock. No apology necessary. Thank you for all your work. Keep it up!

36 sats \ 0 replies \ @Fiat_Revelation 29 Nov

Yes, agreed. I saw in the saloon thread there were others with the same slowness issue and I could wait patiently knowing k00b would come to the rescue and give a prompt update. Pretty decent UX.

286 sats \ 0 replies \ @DarthCoin 29 Nov

Tech shit could happen all the time. IT guys knows that. Understandable and we appreciate your effort to put it back in time. No need to apologize, man. This show us that SN is getting serious. Happy that is not more serious.

192 sats \ 0 replies \ @nerd2ninja 29 Nov

This is a reminder that SN runs on only 1 server. In before SN is yuge

85 sats \ 0 replies \ @WeAreAllSatoshi 29 Nov

I was wondering if it was only me. Thank you for the debrief!

85 sats \ 0 replies \ @denlillaapan 29 Nov

All goody sir, appreciate the transparency and dedication to us compulsory stackers!

75 sats \ 1 reply \ @Rothbardian_fanatic 29 Nov

So often the rebooting or recycling seems to fix vexing problems. I have also found that sometimes the simplest correction to be the one that works. You must be using PostGre. Sometimes a beast. i have had good luck with MySQL before.

10 sats \ 0 replies \ @Rothbardian_fanatic 29 Nov

You could try duplicating your database files, setting up another area and run MySQL on the data to see which runs it better for your usage. Then make the new data fork into both databases and run them in parallel to see which is better.

136 sats \ 0 replies \ @siggy47 29 Nov

Glad to hear you used my favorite technique for solving puzzling technical issues.

67 sats \ 2 replies \ @Public_N_M_E 30 Nov

Show me Devs who respond? Devs that come back to you with an explanation of the problem and what they did to fix it?

Anyone? No? Yeah didn't think so.

This is why these guys are diamonds.

Y'all can have a brofist from Pub 👊

22 sats \ 1 reply \ @ek 30 Nov

Devs that come back to you with an explanation of the problem and what they did to fix it?

To be fair, I would say it looks like most devs don't do that since there's no easy way for them to reach most of their users and vice versa. We're in a unique position where what we're building is a very natural way to interact with our users and for them to interact with us.

You can hear more about this in this panel at minute 17:

Keyan's cheat code is that he built the ultimate focus group.

The feedback loop is incredible on Stacker News. You built the platform that you get users' feedback about the platform

I think there's something generic there, your product should have a mechanism for your users to be able to nag the crap out of you all the time because that's telling you about the reality of them

For full context, watch from 15:28.