Infrastructure

While transferring test data from production to the sandbox-2 environment I noticed failures for a group of pids.
I'll use an example to illustrate (doi_10.5066_F71C1TV7)
https://cn.dataone.org/cn/v2/meta/doi_10.5066_F71C1TV7
CN.SystemMetadata reports checksum as:
<pre>
<checksum algorithm="MD5">46178da6192263921eb755940d716725</checksum>
</pre>

Whereas calculating it from disk gives this:
<pre>
MD5(/var/metacat/documents/autogen.2013062508395355978.1)= efc11787f789b45db29999fb4bd8d745
</pre>

The byte size is also off.
<pre>
<size>16739</size>
</pre>
On disk:
<pre>
-rw-r--r-- 1 tomcat7 tomcat7 16529 Jun 25 2013 /var/metacat/documents/autogen.2013062508395355978.1
</pre>

There are ~70 similar pids that have issues (perhaps more) from our test corpus. They are from the now defunct USGS MN.

I'm not sure what our strategy is since the original MN is not online any longer so we cannot get the "original" bytes from that.