question on resize for large hard disks

[markvtc]

When resizing from the default 50MB to 750GB on a LAMP template, how long should it take - and if the connection to the internet goes kaput and you reconnect, is it still happening in the background or do you have to do it all over again?.

[PeterNic]

The typical speed of disk I/O is about 40MB/sec; so a copy may give you effective 20-30MB/sec (since you pass the traffic twice -- once for read and once for write). See various posts for the disk performance.

Keep in mind that resize is not a block-level copy but a file-level copy. So, it is predicated on the number of files as much as on the size to be copied. Since you are resizing from 50GB, then the max amount of data you have to transfer is 50GB -- I would expect between 30 minutes and an hour to do this resize. It may take longer if you have lots of small files (e.g., hundreds of thousands). What are you seeing?

With regard to connection drop -- usually the connections are quite stable. If you need to survive a connection drop, the easiest solution is to do the resize (or any long operation) not from your client machine by from a GSC/VPS/VDS(*) instance running on the same grid. You create an instance of VDS on the grid, then log into it, maybe run screen or detach the operation from the console. You will need to use ssh login and have your key (either on the VDS or via agent forwarding, at least during the login to the grid). For further convenience, you can use the client-side macros shown in the documentation (in the Accessing your grid topic and/or the end of the shell reference).

Does that help?

Regards,
-- Peter