Check the checksum the second time once we got all checksum values

[kyang2014]

The work describes in this ticket should get started after ticket #315.
After we got sha256 for all BF files and ensure those files are at nearline, it is desirable to do the following:

1. pull BF files in segments(500-1000 orbit each tiem) from near-line to
   a) calculate the checksum again.
   b) ncdump -h to see if there are any corruption.
   c) compare the new checksum with the old one.
2. if there are issues, need to do double-check those orbits and possibly re-generate BF files.

[LandonTClipp]

Sean has stated he is able to provide checksums on NL and is ready to begin the process. His process will take ~5 weeks for the 27972 files that were not initially checksummed on BW. I am actually almost done with these 27972 checksums on BW. We will use Sean's checksums as cross-validation with what I have done.

Note: We will still want to pull the remaining BF files from NL that Sean is not checksumming during the step described in this ticket.

[kyang2014]

Fully agree. From: Landon Clipp [mailto:notifications@github.com] Sent: Friday, April 20, 2018 3:08 PM To: TerraFusion/basicFusion Cc: Kent Yang; Author Subject: Re: [TerraFusion/basicFusion] Check the checksum the second time once we got all checksum values (#318) Sean has stated he is able to provide checksums on NL and is ready to begin the process. His process will take ~5 weeks for the 27972 files that were not initially checksummed on BW. I am actually almost done with these 27972 checksums on BW. We will use Sean's checksums as cross-validation with what I have done. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub<#318 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AJyiE3jzsq96fsKNA-2_jROOyGNTCUUEks5tqkAqgaJpZM4TQw4p>.

[kyang2014]

@LandonTClipp
You've gotten all the sha256sum calculated #315. Hopefully you will work on this ticket so soon.

[LandonTClipp]

We've received some partial results from Sean's checksumming. A manual check of a few orbits returns no difference in hash values, so this seems to be good so far.