I am realizing that the way we are currently querying for ereports by serial is a bit wrong. V1 serials are only guaranteed to be unique within a part number, so we may see (for example) a sled and switch with the same serial. Which means we may get surprising results from an ereport query --- if the user wanted to get ereports from 0XV1:913-0000019:4:BRM42000069 (a Gimlet) and the rack also contained OXV1:913-0000003:2:BRM42000069 (a PSC), they might be somewhat surprised to discover that the bundle they requested of the Gimlet's ereports also has a bunch of "rectifier inserted and removed here" ereports from the PSC. So, we should probably change the way we query ereports by serial to also include the CPN. Sigh. This one's my fault.
Originally posted by @hawkw in #10108 (comment)
I am realizing that the way we are currently querying for ereports by serial is a bit wrong. V1 serials are only guaranteed to be unique within a part number, so we may see (for example) a sled and switch with the same serial. Which means we may get surprising results from an ereport query --- if the user wanted to get ereports from
0XV1:913-0000019:4:BRM42000069(a Gimlet) and the rack also containedOXV1:913-0000003:2:BRM42000069(a PSC), they might be somewhat surprised to discover that the bundle they requested of the Gimlet's ereports also has a bunch of "rectifier inserted and removed here" ereports from the PSC. So, we should probably change the way we query ereports by serial to also include the CPN. Sigh. This one's my fault.Originally posted by @hawkw in #10108 (comment)