Due to an influx of spam, we have had to impose restrictions on new accounts. Please see this wiki page for instructions on how to get full permissions. Sorry for the inconvenience.
The migration is almost done, at least the rest should happen in the background. There are still a few technical difference between the old cluster and the new ones, and they are summarized in this issue. Please pay attention to the TL:DR at the end of the comment.
As mentioned in the bug below, the user is trying to migrate QEMU and it is failing on the unserialization of usbredirparser at the target host. The user does not have USB attached to the VM at all.
I've added a test that shows that serialization is currently broken. It fails at the 'pristine' check in usbredirparser_unserialize().
This check was added with e37d86c2 "Skip empty write buffers when unserializing parser" and restricted further with 186c4c79 "Avoid memory leak from ill-formatted serialization data"
The issue here is that usbredirparser's initialization sets some fields and thus it isn't guaranteed to be pristine.
The parser's basic data is:
| field | value |
|---|---|
| write_buf_count | 1 |
| write_buf | 0xbc03e0 |
| write_buf_total_size | 80 |
| data | (nil) |
| header_read | 0 |
| type_header_read | 0 |
| data_read | 0 |
The current fix is to to ignore write_buf checks as, again, they are not guaranteed to be pristine. usbredirparser library should properly overwrite them when unserializing the data and if there were pending buffers, they should be freed.
Related: https://bugzilla.redhat.com/show_bug.cgi?id=2096008
Signed-off-by: Victor Toso victortoso@redhat.com