On 07/27/2014 11:45 PM, Vladislav Prodan wrote: > Say me, please, how I can transfer between servers ZFS-partitions larger than > 20Gb. > I have 4 partitions with backup data. There are a lot of small pictures. > > NAME USED AVAIL REFER MOUNTPOINT > tank2 1,42T 1,32T 118M legacy > ... > tank2/ay-files 31,9G 1,32T 31,8G > /backup/beastie/upload_ftp/ay/files > tank2/XXXXXmarket-images 135G 1,32T 135G > /backup/beastie/upload_ftp/XXXXXmarket/images > tank2/XXXXXmarket-uploads 17,6G 1,32T 17,5G > /backup/beastie/upload_ftp/XXXXXmarket/uploads > tank2/YYYYmarket 24,4G 1,32T 152K > /backup/beastie/upload_ftp/YYYYmarket > tank2/YYYYmarket/uploads 24,4G 1,32T 24,2G > /backup/beastie/upload_ftp/YYYYmarket/uploads > ... > > Say me, please, how I can transfer it at another server with ZFS. > Servers are connected through unmanaged 100M switch. > Sending via (zfs send) and (ssh host2 zfs recv) ends with error after 20GB of > data. > Sending via rsync is very lasting and requires often restart of rsync. > > What an alternative?
For the initial 20G upload, you could zfs send to a file, and transfer that file via rsync or other resumable file transfer system that ensures integrity : host1# zfs send tank2/XXXXXmarket-images@snapshot > tank2-xxxxxmarket-ima...@snapshot.zfs host1# rsync tank2-xxxxxmarket-ima...@snapshot.zfs host2: host2# cat tank2-xxxxxmarket-ima...@snapshot.zfs | zfs recv tank3 Then the rest can be done via small snapshots and plain zfs send | ssh host2 zfs recv. You could even add bzip2 compression and decompression at each endpoint. This is what I am doing to replicate data of often updated filesystems (mailbox filesystems, with LOTS of small files) between ZFS servers. Cheers, > -- > Stephane LAPIE, EPITA SRS, Promo 2005 > "Even when they have digital readouts, I can't understand them." > --MegaTokyo > > -- > Stephane LAPIE, EPITA SRS, Promo 2005 > "Even when they have digital readouts, I can't understand them." > --MegaTokyo
signature.asc
Description: OpenPGP digital signature
