[DRBD-user] Avoiding semi-fullsync's?

Leroy van Logchem leroy.vanlogchem at wldelft.nl
Wed Apr 27 15:21:59 CEST 2005

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


>>Based on the story below, why are so many bits set
>>while it was perfectly in sync a few moments before
>>the Local IO failure? I really like to aviod these
>>semi-fullsync's.
>>
>>13:28:34 drbd0: PARTNER DISKLESS
>>    
>>
>drbd initiates a full restore in response to that IIRC, because it
>assumes the full remote image has been "lost".
>
Ah thanks. May I conclude that this can be overcome by changing
the on-io-error handler from detach to pass_on or even panic? I'am
still looking for a properly working software method of quickly
leaving and rebooting. Currently I use reboot -n -f , which doesnt
always work - but it does stop heartbeat and drbd quickly.

--Leroy




More information about the drbd-user mailing list