On 08/02/13 11:02, Chris Murphy wrote: > > On Feb 7, 2013, at 4:48 PM, Chris Murphy <lists@xxxxxxxxxxxxxxxxx> > wrote: >> >> ***On Feb 7, 2013, at 6:08 AM, Adam Goryachev >> <mailinglists@xxxxxxxxxxxxxxxxxxxxxx> wrote: >> >>> Basically, on occasion, when a user copies a large file from disk >>> to disk, > >> Why did you reject NFS … > > A sign of brain deficiency, responding to myself… > > Maybe this is an array to array transfer (?) and not LV to LV > transfer where both LVs are on this SSD array? If it's the latter, > geez, use NFS. This transaction is reduced to a rename (i.e. a move). > Super fast. > > Even if it's the former, with different block devices on the same > server, I'm pretty sure NFSv4 can reduce this to a cp between > devices, rather than piping the data all the way to the client, just > to have the client pipe it back to the server. Although I'm sure > someone will correct me if I'm wrong. OK, so again, currently I have: RAID5 DRBD LVM2 iSCSI On the remote machine... iSCSI connects to server and presents block device /dev/sdX Xen which passes through the block device to domU (Windows) disk partition partition is formatted NTFS Now, alternatives to the above involving NFS I imagine are: RAID5 DRBD ext4 (or whatever filesystem format) to create a large file NFS (export the large files to the xen physical machine On the remote machine.... NFS mount loop to present the NFS file as a block device Xen which passes through the block device to domU (Windows) disk partition partition is formatted NTFS LVM offers some advantages: snapshots (disabled to try to improve performance, but in theory...) easy to increase allocated space to a single VM Keeps the entire system as simple block devices. There is only one filesystem (NTFS), everything else is just block devices translated through from layer to lower layer. I'm not sure, but it was my understanding that using block devices was the most efficient way to do this.... Potentially, if the block devices are mis-aligned in some way, I assume this could result in some performance loss, but again I'm not sure it should be this dramatic. Any comments? Thanks, Adam -- Adam Goryachev Website Managers www.websitemanagers.com.au -- To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
