AFP- Accessing a time machine sparse bundle via AFP causes a kernel panic
I have a reproducible kernel panic issue that occurs when I attempt to access a Time Machine sparse bundle over afp. The error exist in both the 9.10-Stable as well. I have included the crash dump and am not sure what additional debugging I should perform.
#2 Updated by Chris Torek over 4 years ago
- Category changed from 78 to 200
- Assignee changed from Chris Torek to Alexander Motin
The crash is in zfs, in (inlined)
ddt_sync_entry (line 1054 reads:
VERIFY(ddt_object_update(ddt, ntype, nclass, dde, tx) == 0);
) There is only one kind of ddt object, a ZAP entry, so clearly
ddt_zap_updatehas returned an error.
The only way for it to return an error is for
zap_update_uint64 to return an error but there are a bunch of paths here that can fail:
fzap_update can all return errors. (In this case we can eliminate
zap_name_alloc_uint64 since it uses
mav is way more familiar with this code than I am, so I'll hand this off to him at this point :-)
#5 Updated by Jordan Hubbard over 4 years ago
This assert going off looks rather like a corrupted pool (does this system have ECC memory?). Suggest full backup of data and recreation of pool. Also check SMART status of drives, while you're at it (unlikely to be the issue here but might as well be comprehensive).