[DRBD-user] Data corrupted after relocation of DRBD-resources

[Roland Kammerer]

On Tue, Nov 19, 2019 at 10:45:31PM +0100, kvaps wrote:
> Hi, today we faced with the weird issue during reloaction some of
> drbd-resources. Everywhere steps were the same: each resources had two
> replicase, one replica was removed, then another one was created on another
> node.
> 
> Almost all of them started having data coruption. Hex dump of two similar
> drives shows zeroes on place where should be data.
> 
> We're using LVM_THIN driver with LINSTOR.
> 
> Kernel version: 4.15.18-18-pve
> 
> DRBD module version:
> 
> version: 9.0.19-1 (api:2/proto:86-115)
> GIT-hash: 6f5fa5d348a99e5eeb09d83c49853d72e614fd07 build by @eb2143fc2fef,
> 2019-07-26 11:40:44
> Transports (api:16): tcp (9.0.19-1)

Does it reproduce with a current version of DRBD? Yours is 2 versions
behind.

Best, rck