Hi,

I'm using diskpacked storage engine, as I don't want to pay the price of 
lot of small files.
Blobpacked came, and I've measured that is quite good - and with the new 
"union" storage, I can even move its localdisk files into a diskpacked 
store.

But blobpacked packs blobs of the same file, so the blobs should be copied 
into the blobpacked store in "natural" order, not lexicographical, as sync 
does.
That's what BlobStreamer interface is for. But that interface does not have 
a HTTP endpoint, as BlobEnumerator (sync) has.

I've tried to implement it, but it is a big undertaking, and I only want to 
shovel my blobs from my diskpacked into my blobpacked.

The I thought maybe camput could get a "rawdpstream" subcommand, and "cat 
diskpacked/pack-*.blobs | camput rawdpstream",
but camput is too intelligent, and I cannot specify the destination exactly 
- it only accepts the server as the server prefix,
contrary to camtool sync, which can.

Shall I try to add a "stream-diskpacked" subcommand to camtool?

Or any other idea?

Thanks,
Tamás Gulácsi

-- 
You received this message because you are subscribed to the Google Groups 
"Camlistore" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
For more options, visit https://groups.google.com/d/optout.

Reply via email to