Hi,
a customer lost 5 OSDs at the same time (and replaced them with new disks before we could
do anything…). 4 PGs were incomplete but could be repaired with ceph-objectstore-tool. The
cluster itself is healthy again.
Now some RBDs are missing. They are still listed in the rbd_directory object but cannot be
access with e.g. rbd info.
rbd ls says: no such file or directory
I assume that is because the associated rbd_id.name object has been lost.
I tried to create this object by means of "rados put" but still no luck:
2020-08-13 11:47:27.557 7fa1e2ffd700 -1 librbd::image::OpenRequest: failed to retrieve
image id: (5) Input/output error
2020-08-13 11:47:27.557 7fa1e27fc700 -1 librbd::ImageState: 0x55b9ebb75b10 failed to open
image: (5) Input/output error
2020-08-13 11:47:27.557 7fa1e27fc700 -1 librbd::io::AioCompletion: 0x55b9ebadd3a0 fail:
(5) Input/output error
rbd: error opening vm-501-disk-2: (5) Input/output error
What are the options to repair an RBD pool so that at least the
RBDs (most data objects are still there) are available again.
Regards
--
Robert Sander
Heinlein Support GmbH
Schwedter Str. 8/9b, 10119 Berlin
http://www.heinlein-support.de
Tel: 030 / 405051-43
Fax: 030 / 405051-19
Zwangsangaben lt. §35a GmbHG:
HRB 93818 B / Amtsgericht Berlin-Charlottenburg,
Geschäftsführer: Peer Heinlein -- Sitz: Berlin