pt-archiver not using --columns specified

I have a table with large blobs in it that I’m trying to move to a new server. I want to move all the records without the blobs first, then I’ll run pt-table-sync later to update the blobs. I used the --columns option in pt-archiver to specify every column except the blob one I want to ignore. However, pt-archiver picked 3 of the columns and didn’t use the other ones I specified.

The command I’m using is:

pt-archiver --progress 1000 --columns=id,document_id,loadDate,fullName,fileName,fileExtension,type,loadUser,sizeMB --database=db --source … --dest … --no-delete --why-quit --txn-size 500 --where “1=1”

The dry-run shows:

INSERT INTO db.attachments(id,document_id,type) VALUES (?,?,?)

Why did it only pick these 3 columns? I told it all the columns I wanted…

Even I am facing same problem, I think it’s because of column names, it has caps in between the words (fullName) may be pt-archvier doesn’t support it, not sure how to ignore it .

Could you share the output of



Hi again,

Internally, pt-table-sync is converting all field names to lowercase.

--columns action,minlat,minlon,ipaddr



I have specified field names in lower case it worked.