[DRBD-user] Block out of sync right after full resync

Lionel Sausin ls at numerigraphe.com
Mon Oct 13 10:18:11 CEST 2014

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.

Le 12/10/2014 23:43, aurelien panizza a écrit :
> Do you suggest me I should disable the swap on my server ?
Not at all. I suggest you either:
- (easy): ignore this probably benign message
- or (harder): find some other way to assert the validity of your 
secondary node, because, by design, DRBD can't do it 100%.

> I'm using EXT4
> How can I check if my primary and my secondary are synchronized using 
> a command like fsck ? (online)
You can't.
You'll need a downtime - the other node must become primary to access 
the data.
It's going to be quite time consuming.
You need to have a very clear picture of all the layers involved and how 
they access the data.
Then, on a quiet day, unmount cleanly on the primary node, swap 
primary/secondary roles, test your data on the other node "somehow" - 
maybe fsck, maybe comparing md5sums, maybe comparing with backups - be 
creative, only you can decide how to do this in a meaningful way.
If you do detect errors there, maybe you have a hardware problem worth 
investigating, or a very subtle and nasty bug somewhere .
Otherwise, then just ignore the messages and live happily ever after.
> I found explanation about "data getting modifying in flight" but no 
> "workaround" or everything on how I can avoid getting out of sync block.
There's no solution, and hopefully there's no problem in the first place.


More information about the drbd-user mailing list