Re: FYI (large devices = looooonnngg processing times)

2013-05-01 Thread Bill McGonigle
On 04/27/2013 08:06 PM, Curt Howland wrote: dd if=/dev/random of=/dev/sdb or whatever it was, USB3. 6 days later You might have just run out of entropy on the PRNG. That and small writes will kill you. Try: dd if=/dev/urandom of=/dev/sdX bs=2M next time. It should finish in

Re: FYI (large devices = looooonnngg processing times)

2013-04-30 Thread Tom Buskey
ZFS added RAIDZ3 (triple parity) was because the likelihood of hitting another error before a resilver finishes is likely with 3TB and current ECC on drives today. If you're using 4 TB drives, you should be using double parity (RAID6 or 3 way mirrors). On Sat, Apr 27, 2013 at 7:50 PM,

Re: FYI (large devices = looooonnngg processing times)

2013-04-27 Thread Michael ODonnell
Even on just a silly little RAID1 mirror on a multi-Tb array I dread seeing the various messages announcing routine maintenance and diagnostic operations as they take forever and don't come for free, resource-wise... ___ gnhlug-discuss mailing list

Re: FYI (large devices = looooonnngg processing times)

2013-04-27 Thread Curt Howland
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On Saturday 27 April 2013, Michael ODonnell was heard to say: Even on just a silly little RAID1 mirror on a multi-Tb array I dread seeing the various messages announcing routine maintenance and diagnostic operations as they take forever and

Re: FYI (large devices = looooonnngg processing times)

2013-04-27 Thread Bruce Dawson
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 I remember back when the Digital TurboLaser systems came out. At the time they were enormous, and my job was to develop testing tools for them. Management just about fell off their chairs when I told them it would take a *week* (running 24x7) just to