Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Lost all datas after DELETE command #8822

Closed
max3163 opened this issue Sep 13, 2017 · 7 comments · Fixed by #8981
Closed

Lost all datas after DELETE command #8822

max3163 opened this issue Sep 13, 2017 · 7 comments · Fixed by #8981

Comments

@max3163
Copy link

max3163 commented Sep 13, 2017

Bug report

Influxdb version = 1.3.5
On linux go1.8.3
Intel(R) Xeon(R) CPU E5-2620 with 64Go of RAM

Sometimes, when we try to delete somes points on a specific measurement , we lost 2 or 3 days of datas in our database. ( all measurements are impacted ! )

Exemple of DELETE query :

DELETE FROM events WHERE application = 'cosmos' AND time < 1505223439180000001 AND time > 1505223439179999999

Same issue with :

DELETE FROM events WHERE application = 'cosmos' AND time = 1505223439180000000

You can find a extract log when the issue appear with many messages about full compaction aborted...
If you need a more details logs , tell me.
We have many application which write data in parallel.

delete.zip

@e-dard
Copy link
Contributor

e-dard commented Sep 15, 2017

@max3163 are you saying that you have points in the events metric that are outside the time window [1505223439180000001, 1505223439179999999] and are deleted?

@max3163
Copy link
Author

max3163 commented Sep 15, 2017

Not really, I'm saying that I lost most of the time last 3 days of datas points in all my measurements present in the same database that the 'events' metric.

@e-dard
Copy link
Contributor

e-dard commented Sep 16, 2017

So you're saying that within the specified time window you have lost points that belong to a series that doesn't have the application=cosmos tag pair?

@max3163
Copy link
Author

max3163 commented Sep 16, 2017

Yes exactly !
We lost all datas points on the database 'qualif' for a period of time of about 3 days.
All tag pair are impacted and some measurements don't have the application tag at all.

@jwilder
Copy link
Contributor

jwilder commented Oct 19, 2017

This should be fixed by #8981.

@max3163
Copy link
Author

max3163 commented Nov 14, 2017

I've test the latest version of Influxb with your fix but I still have the issue.
See in attachment a new log file, the issue appear when I start delete 10 points in a row at '2017-11-13T16:16:22Z'

After that, one week of data is lost in all the database.
The issue appear in our backup server with no writer just some queries from grafana.

log2.txt

Can you re-open the bug ? Or I need to create a new one ?

@daviskirk
Copy link

Same here!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging a pull request may close this issue.

5 participants