ZFS send/recv with ashift 9->12 leads to data corruption #12770
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: Paul Dagnelie <pcd@delphix.com>
|
As an experiment, I did the following with this PR applied (and I'm reasonably confident I wasn't running the wrong module...): Attempting the send a second time to capture the send stream for you (with the same command with | tee /tmp/sendstream in the middle) panicked the VM: So that's new and exciting. (I was just looking to see how raidz played in the sandbox with p/lsize here...) I can't reproduce it on VM reboot though... (I rebooted the VM before running this in the first place, though, so ???) I'm going to go memtest just to be sure nothing is broken, since this system is reasonably new, but I haven't had any issues before this... edit: It reproduces if I freshly recreate the pools and write the data anew, but not on reboot. How very odd... |
|
According to zstreamdump on reproducing it... (If you'd like, the "invalid checksum" send output is here.) (FYI, I just booted this patch on my non-VM testbed, and it panics the same way. So I don't think there's much chance of it being just my new system being flaky...) |
ahrens
reviewed
Nov 16, 2021
|
I had a long writeup of what's going wrong, but since GH won't let you comment on unchanged lines: I think the issue is that, as-is, your code doesn't ever set srdp->io_compressed if the arc_read call succeeds, so you end up if the ARC has the data treating it as uncompressed, and Bad Times Ensue. I would suggest setting it outside of the if, at which point, I don't think you need the one inside the if any more? Haven't tested it, yet, but it seems correct from what I observed (if it's in the ARC, you get into dmu_dump_write with io_compressed=0). (Edit: Quickly testing with just a |
Signed-off-by: Paul Dagnelie <pcd@delphix.com>
|
This new commit should address the issue; I also tested the patch with sends hitting and missing in the ARC. |
ahrens
approved these changes
Nov 16, 2021
rincebrain
approved these changes
Nov 16, 2021
tonyhutter
pushed a commit
to tonyhutter/zfs
that referenced
this pull request
Dec 8, 2021
Improve the ability of zfs send to determine if a block is compressed or not by using information contained in the blkptr. Reviewed-by: Rich Ercolani <rincebrain@gmail.com> Reviewed-by: Matthew Ahrens <matthew.ahrens@delphix.com> Signed-off-by: Paul Dagnelie <pcd@delphix.com> Closes openzfs#12770
tonyhutter
pushed a commit
to tonyhutter/zfs
that referenced
this pull request
Dec 8, 2021
Improve the ability of zfs send to determine if a block is compressed or not by using information contained in the blkptr. Reviewed-by: Rich Ercolani <rincebrain@gmail.com> Reviewed-by: Matthew Ahrens <matthew.ahrens@delphix.com> Signed-off-by: Paul Dagnelie <pcd@delphix.com> Closes openzfs#12770
snajpa
pushed a commit
to vpsfreecz/zfs
that referenced
this pull request
Feb 6, 2022
Improve the ability of zfs send to determine if a block is compressed or not by using information contained in the blkptr. Reviewed-by: Rich Ercolani <rincebrain@gmail.com> Reviewed-by: Matthew Ahrens <matthew.ahrens@delphix.com> Signed-off-by: Paul Dagnelie <pcd@delphix.com> Closes openzfs#12770
nicman23
pushed a commit
to nicman23/zfs
that referenced
this pull request
Aug 22, 2022
Improve the ability of zfs send to determine if a block is compressed or not by using information contained in the blkptr. Reviewed-by: Rich Ercolani <rincebrain@gmail.com> Reviewed-by: Matthew Ahrens <matthew.ahrens@delphix.com> Signed-off-by: Paul Dagnelie <pcd@delphix.com> Closes openzfs#12770
nicman23
pushed a commit
to nicman23/zfs
that referenced
this pull request
Aug 22, 2022
Improve the ability of zfs send to determine if a block is compressed or not by using information contained in the blkptr. Reviewed-by: Rich Ercolani <rincebrain@gmail.com> Reviewed-by: Matthew Ahrens <matthew.ahrens@delphix.com> Signed-off-by: Paul Dagnelie <pcd@delphix.com> Closes openzfs#12770
snajpa
pushed a commit
to vpsfreecz/zfs
that referenced
this pull request
Sep 18, 2022
Improve the ability of zfs send to determine if a block is compressed or not by using information contained in the blkptr. Reviewed-by: Rich Ercolani <rincebrain@gmail.com> Reviewed-by: Matthew Ahrens <matthew.ahrens@delphix.com> Signed-off-by: Paul Dagnelie <pcd@delphix.com> Closes openzfs#12770
snajpa
pushed a commit
to vpsfreecz/zfs
that referenced
this pull request
Sep 19, 2022
Improve the ability of zfs send to determine if a block is compressed or not by using information contained in the blkptr. Reviewed-by: Rich Ercolani <rincebrain@gmail.com> Reviewed-by: Matthew Ahrens <matthew.ahrens@delphix.com> Signed-off-by: Paul Dagnelie <pcd@delphix.com> Closes openzfs#12770
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Motivation and Context
See #12762 for details.
Description
We improve the facility of zfs send to determine if a block is compressed or not by using the information contained in the blkptr, rather than relying on an lsize vs psize difference (which is usually, but not always, correct). Also removed an erroneous assertion in the arc codepath.
How Has This Been Tested?
Replicated the setup in the initial issue and verified that
dozer/test-secondandtank/testcontained the same contents after the fix was applied.Types of changes
Checklist:
Signed-off-by.