Anyone familiar with any of these three problems:
1) I create a volume with 4 bricks (gets performance X), then add four
bricks (gets performance X/2)
This does not make a lot of sense to me. I understand there could be
a bottleneck, but why would the performance *decrease* (and by exactly
half)?
2) rsync -av /somedir/foo /gfs/otherdir/
Sometimes works reasonably well. Other times never finishes. When it
did not finish, the directory was larger and more nested
3) gluster migrate operation fails with a socket error in the logs.
Seems like a name resolution problem or something
"reading from socket failed. Error (Transport endpoint is not
connected), peer (127.0.0.1:1021)"
Why is it trying the loopback? That seems like a bad choice. The
volume works fine. Just the migrate operation fails. I've seen this
question asked in the archives, but have not seen a good answer.
Thanks
Chuck
