NASLite Network Attached Storage

i know little about how Naslite boots but on older systems with slow cpu's and little memory boot times can seem an age - not a problem when a system is on 24/7 but a couple of people who have looked at my system want to have quick access from a standing start

as i understand it Naslite clean boots each bootup - decompressing and looking at the hardware and selecting and installing the correct drivers for the system it is on - all that takes time

is there - could there be a way to save the image in memory and have that installed rather than start clean every time - the system - hardware etc is the same as last time it booted

i have no idea if that would speed up boot time - its just that when asked - my answer fell very much into - i have no idea

Tony

I have just timed mine doing a remote reboot and it took just under 2 minutes to do that reboot from me issuing the y confirm to being able to telnet back in. I will do a boot from power off and time that for you if it would help.

from powered down it takes 1 min 41 secs to be at the login prompt.

P.S. my system is on a CF card on an IDE adapter are you booting from the CD each time I wonder?

Doug

hi Doug - speed of bootup is something that will never satisfy some - my PIII - booting from a hard drive is up and running in just over a minute - i am more than happy with that

i know it can vary depending on method of bootup and spec of hardware

i asked the question as i had been asked it by a couple of people and i had no idea as to the answer - i suspect it is no!

The answer could be "yes", but would require quite a bit of development effort on the part of SE. So the most likely answer, for now, would be "no".

Tell your questioners to get a cup of coffee after pushing the power button

There are versions of Linux that have been ported to a BIOS ROM and boot in a little more than 10 seconds from cold. That said, there is little practical reason for an appliance such as NL to boot fast. Most SAN and NAS devices are starting to employ a staged power down of unused drives after a set amount of idle time. This saves a good chunk of power and gives one the fast response you are looking for. The problem is that Linux would not like having a volume go down and then up like that. It would require that the volume be unmounted and then the drive spun down. When the drive was hit again it would require that the drive be spun back up and then mounted again. In addition, it is allot more ware on a drive to spin it up and then down than to just leave it running. I have two drives that have 33,000 spin hours on them each. They have run without a hitch for the better part of 4 years.

Personally I see little to no need for this feature. I would prefer that they focus their attentions on software like SANLite with iSCSI and fiber channel targets.

Mike

i do not disagree with your general argument Mike - but each to their own - the software is used at many levels - the discussion on the raid problems would seem to be one [extreme] end of its use

the questions i was asked is probarbly the other end - just a simple single drive system that can be switched on via wol on demand and powered down when finished with - i agree 1 minute to wait is not long but on slow pll with 128meg of ram it would be longer than that

georg's idea is best - coffee is always an answer

Hard to argue against the single drive and once in a while thing for the need for a device to power down. I can see the utility of it. Hard on the drive though.

Mike