[DRBD-user] Avoid blocking IO when node fails

Lars Marowsky-Bree lmb at suse.com
Thu Feb 13 16:50:09 CET 2014

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

On 2014-02-13T11:09:14, Roberto Munoz Gomez <roberto.munoz.gomez at diagroup.com> wrote:

> The problem is that while the other node is being fenced, the active one cannot access the DRBD and all the IO calls are blocked until the fenced node comes alive again. And that is what I want to avoid. I want the survival node to continue working.

This is not possible. Until the fence has completed, it is not clear
which side will actually win, so both sides freeze - to avoid data
corruption after a partitioned network.

If you want that, split the data into independent slices that are not
active/active on DRBD.


Architect Storage/HA
SUSE LINUX Products GmbH, GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 21284 (AG Nürnberg)
"Experience is the name everyone gives to their mistakes." -- Oscar Wilde

More information about the drbd-user mailing list