About today’s outage [Update May 30]
You can also read this post in Chinese.
UPDATE: We have completely recovered all data. Read more here.
Today we had our first major database crash and to tell you that we were fully prepared will be a lie. We weren’t. Due to human error during regular maintenance work, we’ve simultaneously lost two database servers (master and slave) and their data. Backups are in place, but due to the huge amounts of data and some other issues, it will take time for restoration.
What is affected:
- Articles older than 3 days will not be available at the moment. This includes saved web pages and mail2tag emails.
- Some feeds might not update for a while, because we are now ignoring articles that doesn’t have valid pubDate field. This is a temporary precaution to avoid duplicating content.
- Your credentials, including linked social accounts
- Your preferences, including per-feed settings.
- Folders
- Subscriptions, including Facebook, Google+, and VK feeds
- Rules
- Bundles
- OPML subscriptions
- Monitoring feeds
- Teams
- Highlighters
- Email contacts
- Reading statistics
We will update this post when we have more information on the recovery.
UPDATE May 28 23:30 GMT:
ETA for data recovery first results is June 01. We expect to have all articles between 2005 and August 2015 to be available by then, including all starred and tagged ones. We will then proceed with newer articles.
We will also temporarily extend the 30-day limit for unread articles to 45 days, so you will be able to catch up with your reading.
UPDATE May 30 17:00 GMT:
About the promised compensations – All current subscribers with active rebilling plans will receive coupons for 2 months of Professional plan on their email. All other users that have active Starter/Plus/Professional/Team accounts without active rebilling plans will automatically receive a two-month extension tomorrow.
Your data is safe! See our transparency report from today for details.