Don't fence clients with rid==0 #190
Open
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
Occasionally during the export-lookup-evict-race test we see the following failure dmesg output when a server fences a client that has no valid rid: [ 828.379546] sysfs: cannot create duplicate filename '/fs/scoutfs/f.b928e1.r.7b36c0/fence/0000000000000000' ... [ 828.379773] kobject_add_internal failed for 0000000000000000 with -EEXIST, don't try to register things with the same name in the same directory. [ 828.385946] scoutfs f.b928e1.r.7b36c0 error: client fence returned err -17, shutting down server This fails the test. Fencing these clients is unwanted, but we definitely don't want to create duplicate sysfs entries for it, either. Don't fence clients like this, just return 0. Signed-off-by: Auke Kok <[email protected]>
for reference - caller path is known. |
zabbo
requested changes
Oct 10, 2024
| @@ -238,6 +238,11 @@ int scoutfs_fence_start(struct super_block *sb, u64 rid, __be32 ipv4_addr, int r | |||
| struct pending_fence *fence; | |||
| int ret; | |||
|
|
|||
| if (!rid) { | |||
There was a problem hiding this comment.
Sadly, I don't think this'll work. Depending on where the fence came from we might not end up cleaning it up. See the _fence_next calls in the server that lead to _recov_finished. We'll need to more about the precise source of the fencing request to avoid it.
There was a problem hiding this comment.
So, we know it comes from scoutfs_net_reconn_free_worker+0x1bd and not the other 2 callers. I'll look into that.
Also, we can just stopgap and not create sysfs entries here, too, and just let all the fencing happen. The worst that could happen, I guess, is that we end up fencing the wrong clients that also have rid=0, so maybe not such a big deal?
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.
Occasionally during the export-lookup-evict-race test we see the following failure dmesg output when a server fences a client that has no valid rid:
[ 828.379546] sysfs: cannot create duplicate filename '/fs/scoutfs/f.b928e1.r.7b36c0/fence/0000000000000000' ...
[ 828.379773] kobject_add_internal failed for 0000000000000000 with -EEXIST, don't try to register things with the same name in the same directory. [ 828.385946] scoutfs f.b928e1.r.7b36c0 error: client fence returned err -17, shutting down server
This fails the test. Fencing these clients is unwanted, but we definitely don't want to create duplicate sysfs entries for it, either.
Don't fence clients like this, just return 0.