Details
-
Bug
-
Status: Resolved
-
Minor
-
Resolution: Fixed
-
2.0
-
None
Description
The image retrieval and/or does_image_exist subroutines in xCAT.pm do not verify if the retrieval was successful if SCP runs and returns exit status 0. I have seen an instance where only the .sfdisk and .mbr files were retrieved from another management node. The subroutine reported that the image was retrieved sucessfully.
The subroutine currently loops through its image library partners and stops once it finds one with the image. It may be beneficial to first loop through all management nodes getting information about whether or not they have the image. We can then compare the sizes and use the largest one. After SCP is done, we can recheck the size to make sure they match.
It's possible another management node you attempt to retrieve from doesn't contain a complete copy of the image. This may happen if that management node is also retrieving the same image at the same time, which isn't all that unlikely if a new block request is being processed and several reservations for the same image are made at the same time.