Opened 9 years ago
Last modified 9 years ago
#72 new defect
Self-destructive Salvage Behavior
| Reported by: | Owned by: | Eric Swenson | |
|---|---|---|---|
| Priority: | minor | Milestone: | |
| Component: | General | Version: | |
| Keywords: | copy_dump salvager rpvs | Cc: |
Description
I once did a crash dump but did not do an esd afterwards.
The volume was rebooted with rpvs, but the volume was badly damaged.
I observed that the 'copy dump' step ran before the salvage completed and I hypothesis that this contributed to the poor state of the volume after salvaging.
I see no reason why the 'copy dump' step should not wait for the salvage to complete. The data being copied is in the dump partition and should be completely untouched by the salvager; the copy operation will be adding segments to volume, and that should wait until salvaging is complete to avoid further damage to the volume.
Note:
See TracTickets
for help on using tickets.
While I agree that it might make sense for Utility to wait until salvaging is complete, it would be really difficult (system redesign) to prevent any process from writing to the file system. If the many daemons are launched, they are all writing to the file system as part of their normal operation. The system is supposed to be able to handle that.