Jorg Vervaet

@trakt @ZERO_ISM How will you recover it if you don’t have a backup and no mirror?

@ZERO_ISM We wouldn’t know what is missing. If we knew, we’d recover it.

Trakt website & Plex scrobbler are back online! Trakt API will be back next week. We're closely monitoring the servers while things ramp up. We're deeply sorry for the extended downtime & we'll do better moving forward. Updates to our automated backups are already in place.

@trakt Was it unplanned the data corruption? I think there was no plan to handle any kind of problem. A week down. A backup that is more then a month old? Really? How can you communicate these things and feel ok with it?

On Dec 11 at 7:30 PST, our main database crashed & corrupted some of the data. This is unplanned & we've been working around the clock to restore it. It's a painstakingly slow process given the scale of data. We unfortunately have no ETA but we're doing everything we can.

@trakt is there an other app that I can use instead of trakt because I think this one is not reliable if they even get in online again. Who needs a week for this?

@Anti_ZZ_ @trakt My thoughts exactly and it isn’t even monthly. They really don’t know what they are doing

@trakt You only backup monthly?

Earlier this week, we restored a full database backup from Nov 7 as the starting point. We’ve incrementally imported data from Nov 7 up until the database crashed on Dec 11. There is data corruption, but we’ve recovered a good percentage thus far.

@trakt @great_vc Do you really know what you are doing? It is almost a week. It is not rocket science. Also you have a backup that is more then a month old? Really? What did you think. That you are invulnerable? No real backups no mirrors. So really nothing. So what are you trying to do?

@great_vc The full restore is from Nov 7, so data after that point could be corrupted or lost. Recovering the data after Nov 7 is what we’ve been doing all week and we’ve been able to recover a lot so far. We have more details about this once we’re ready to bring the website back online.

@trakt So you don’t have 2 servers? Always have 2 mirrors so if 1 fails the other 1 takes over. Don’t you have a backup?

@simonw I would use Pyspark

If someone gives you a CSV file with 100,000 rows in it, what tools do you use to start exploring and understanding that data?