blobs, mssql
I am having serious problems handling blobs.
I am using 1.0.21 on a windows server 2003 machine connecting to a locally running instance of MS SQL and migrating to a remote mysql 5 running on a linux machine. My source db has a table with 25,000 binaries (PDF), though i sincerely doubt any one of them is over 1MB.
The problem is as follows:
1. if i use run_migration from the linux side, then everything goes smoothly EXCEPT that many of the blob rows disappear. I suppose this could be the 4MB constraint, although the nice thing to do would be to tell me each row where this occurred, and keep the row, just nullify the blob column. however, there is no way that i know of to lift the 4MB (or any other) blob constraint via run_migration -- the linux tool.
2. plan b was to install the windows version on the windows server. in this method, I use the "BLOB stream" option -- and it seems that checking this box is being interpreted as checking the "ignore blobs" option. The only blob I have in my whole database is the column I described above, and this table gets migrated perfectly with null blob column when I choose BLOB stream. However, when I use BLOB stream, several other tables are created, but their data is just not migrated. without any explanation.
[as an aside: one thing that concerns me about this is that these tables with their data missing have FK relations to many other tables, but mysql does not seem to mind. I see int he creates.sql that the constraints *should* be migrated, so I would EXPECT mysql to complain when it was missing all this referential integrity... but perhaps this is just because we turn off referential integrity to allow the transfer of data, and when it is turned back on, it is assumed to be correct, and not actually checked...]
3. suspecting that the blob stream checkbox is broken, i tried to run the migration with that checkbox unchecked, but checking the IGNORE BLOB checkbox. sure enough, my blobs are NOT ignored, and they are migrated in this configuration. but then i still end up with some of the blob rows missing.
4. so now i am trying to use the "no 4MB restriction" option to see what happens.
what makes this so frustrating is that in my very first try (the run_migration) it all worked GREAT except for the missing rows. the rows that WERE there, were PERFECT. and each of these iterations takes several hours, which also is enough to make anyone crazy.
i searched this entire forum and the bug database, and the help manual, and it does not seem like the blob stuff has been really exercised much. please please please help!
thanks!