Bug #3555
Download/Upload missed data with very big files
Status: | Closed | Start date: | 04/05/2013 | |
---|---|---|---|---|
Priority: | High | Due date: | ||
Assignee: | Stavros Sachtouris | % Done: | 100% |
|
Category: | kamaki | Spent time: | - | |
Target version: | v0.8 |
Description
Still not sure if it is a kamaki bug, but should be reported here. Kamaki seems to misshandle very large file downloads.
Test case:
with an image > 11GiB
download -A <owner-id> image:<image-id> download the file from the UI ls -l both files: are equal in total size md5sum both images: cd3d2e6d09ab829d82c49e3fc8a9b52b kamaki-downloaded.diskdump 88e7cc2615123623c8a4b7b246c7108b ui-downloaded.diskdump Why are they different?
History
#1 Updated by Stavros Sachtouris over 11 years ago
Note: The -A option is not avaiable in 0.7.X, use --account instead
#2 Updated by Stavros Sachtouris over 11 years ago
Nikos Skalkotos reports: the kamaki-downloaded md5 is wrong, the ui md5 is correct
#3 Updated by Stavros Sachtouris over 11 years ago
Turns out upload is not affected by this problem.
#4 Updated by Stavros Sachtouris over 11 years ago
- Status changed from New to Resolved
- Priority changed from Medium to High
Pinned down: When a file contains at least 2 blocks with the same hashtag (== exactly the same content), kamaki downloads only one of these equal blocks and ignores the rest.
#5 Updated by Stavros Sachtouris over 11 years ago
- % Done changed from 0 to 100
#6 Updated by Stavros Sachtouris over 11 years ago
- Status changed from Resolved to Closed
A functional test checks this problem so it is safe to say it is fixed