[OpenIndiana-discuss] zfs incremental send / recv looping forever?

[Gabriele Bulfon]

Hi,
I have a pool that is incrementally sent to a backup pool  every day.
It used to work since last week, now I get this in a loop, and I have to kill the send / recv, or the log will fill up soon:
promoting data-backup/sonicle/zones/webtop/ROOT/zbe
another pass:
promoting data-backup/sonicle/zones/www/ROOT/zbe
another pass:
promoting data-backup/sonicle/zones/webtop/ROOT/zbe
another pass:
promoting data-backup/sonicle/zones/www/ROOT/zbe
another pass:
promoting data-backup/sonicle/zones/webtop/ROOT/zbe
another pass:
promoting data-backup/sonicle/zones/www/ROOT/zbe
another pass:
promoting data-backup/sonicle/zones/webtop/ROOT/zbe
....
....and on and on and on
What is it trying to do?!? What is "another pass"???
The script is as easy as:
zfs send -R -i $snapshot_fromsnap $snapshot_tosnap | /usr/sbin/zfs receive -Fduv $destpool
Any idea??
BTW, I already did a full send from scratch again, and it worked fine, then I got the same error running the incremental send.
I have to check but I think those zones are both cloned from another one.
So probably the problem comes from the clone.
But why loop forever?
Can I feel safe to promote the clones and get rid of the original snap?
Gabriele.