osd: avoid push raw clone when head missing

[Aran85]

if we are missing head, the snaps of clone will not present.
push raw clone will cause peer's SnapMapper cheat.

Fixes: http://tracker.ceph.com/issues/23030

Signed-off-by: Zengran Zhang zhangzengran@sangfor.com.cn

• References tracker ticket
• Updates documentation if necessary
• Includes tests for new functionality or reproducer for bug

[Aran85]

@liewegas would you have a look? thanks.

[xiexingguo]

See https://tracker.ceph.com/issues/24396 (#22396 seems to be a work-around rather than a fixer, though)

[Aran85]

@xiexingguo thanks pointed out, i think there still may be some cases that we can avoid crash or inconsistent, i mean the cases different with the "unexpected clone". like the head object is really missing?

i think Sage and David had decided to tolerate such an error in the hope that a repair can be performed instead of crash peer. is this reliable when we removing snapshot? because the snapmapper will cheat..

[xiexingguo]

i think there still may be some cases that we can avoid crash or inconsistent, i mean the cases different with the "unexpected clone". like the head object is really missing?

In general we should recover the head object first before trying to recover any descendant clone objects...

[Aran85]

In general we should recover the head object first before trying to recover any descendant clone objects...

yes, this is why i think return -EINVAL is ok.. but seems that "unexpected clone" will cause head objecte be deleted and leave the stray clone object...

[jdurgin]

is this obsolete with the assert/fix added in #27575?

[Aran85]

@jdurgin yeah, they are interrelated, but not sure if it is obsolete, maybe the #27575 is not a sufficient fix to avoid push raw clone..

[stale]

This pull request has been automatically marked as stale because it has not had any activity for 60 days. It will be closed if no further activity occurs for another 30 days.
If you are a maintainer or core committer, please follow-up on this pull request to identify what steps should be taken by the author to move this proposed change forward.
If you are the author of this pull request, thank you for your proposed contribution. If you believe this change is still appropriate, please ensure that any feedback has been addressed and ask for a code review.

[neha-ojha]

Let's run this through a rados run to see how it does.

[tchaikov]

• http://pulpito.ceph.com/kchai-2019-09-07_14:44:52-rados-wip-kefu-testing-2019-09-07-0159-distro-basic-mira/
• http://pulpito.ceph.com/kchai-2019-09-08_13:41:17-rados-wip-kefu-testing-2019-09-07-0159-distro-basic-smithi/

[stale]

This pull request has been automatically marked as stale because it has not had any activity for 60 days. It will be closed if no further activity occurs for another 30 days.
If you are a maintainer or core committer, please follow-up on this pull request to identify what steps should be taken by the author to move this proposed change forward.
If you are the author of this pull request, thank you for your proposed contribution. If you believe this change is still appropriate, please ensure that any feedback has been addressed and ask for a code review.

[stale]

This pull request has been automatically closed because there has been no activity for 90 days. Please feel free to reopen this pull request (or open a new one) if the proposed change is still appropriate. Thank you for your contribution!