I’m running backups (both a scheduled task as well as a manual backup) from
GAE to a GCS bucket. The GAE source datastore has about 45 kinds. The
datastore is a small preliminary test store. Both the manual and scheduled
task backups exhibit the following characteristics. I’m interested in
understanding the 2nd part.
The first part of the output in the GCS bucket looks to be (meta)info about
the kind backups: for each kind, there is an object
ah[long ID #].[kindName].backup_info
All of these are of size < 1000B.
The second part of the output in the GCS bucket contains two to four
objects for each kind in the GAE datastore:
datastore_backup_[backup name]_[date]_[kindName]-[ID#]-output-[N]-attempt-1
SIZE
0 =< N =< 3; every kind has at least 2 such objects.
The ID# appears to be the same for all 2nd half objects.
The SIZE of most of the objects is 0B;
Even kinds that I know have no entities in the datastore have at least 2
such objects, though they each have SIZE = 0B.
Most of the kinds in the datastore have just one entity recorded.
For kinds where there exist entities in the datastore, usually there is
only one corresponding object (output-N) where the size is non-zero. For
WHICH N the object is non-zero appears to be non-predicable.
Also, for the kinds which have multiple entities stored (up to 10-15), 2
corresponding objects are both non-zero.
Almost all the SIZE values are multiples of 32KB: 32KB, 64KB, 224KB,
256KB, 352KB, 448KB, 768KB, 1.47MB, 1.75MB
Can someone provide an explanation of the structure of this 2nd half?
Simple tests show that the backup/restore is working. Generally, the
variation in SIZEs of the non-zero objects seems to roughly correspond to
the variation in size of the stored corresponding entities in the
datastore. But all the 0B objects are confusing, as is the somewhat random
mapping on non-zero objects to the output-N entities. If possible, I’d
like to understand.
Another (small) question: what is the relationship (if any) between the
long ID # of the 1st half (backup_info) objects, and the ID# and backup
name for the 2nd half objects?
Thanks in advance,
—Ken Bowen
--
You received this message because you are subscribed to the Google Groups
"Google App Engine" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/google-appengine.
For more options, visit https://groups.google.com/d/optout.
