NASLite Network Attached Storage

My 2 cents. New to NASlite (2 weeks) but have the same problem. Goofed while setting up the UPS shut down script. But HDD not Maxtor but with a 120 WD and a 300 Seagate. They were on different IDE cables (primary and secondary). Both were hooped as far as NASLite was concerned but both can be read fine from XP on external USB drive. Suspect it is not the drive type or cable issue but a (Linux) issue. But not a linux expert. Certainly there must be some repair software to fix the drives without having to reformat and copy. If XP can read them??

Mine has been up for over a year with little problem at all. I would think hardware before software on this one.

Mike

Can't be certain of the long-term position, but certainly my system was fine once I fitted the new HDD and set it up... I didn't alter anything else such as cables... so I think it may be either a "maxtor" incompatibility issue, or a software glich. I'm using NasLite+ v1.1.... am I right in thinking that later versions include a disk-checking utility? (and if so, would this repair, as wek asks?)

Hugo

I've had my system set up for about a week. Over the weekend the NASLite+ box crashed (just disappeared off the network, could not telnet in at all). After resetting the machine one of the 3 drives disappeared. After reinitializing it and rebooting it worked fine but I lost all of the data.

The drive that disappeared was the only drive on the primary channel (Maxtor). Two drives (both Seagate) on the secondary channel were just fine. Just my $0.02.

I notice from other threads on this site that this is a very common problem, power outages or just pulling the plug. It seems to be related to the age/type of system. One person had two systems going when the power went out. All drives hooped on the old system but the new system had no problem.

I found an old 80 Maxtor and tried it and sure enough, pull the plug and it is dead to NASlite as well as the Seagate and WD drives. Good thing everything is easily readable through XP. But for the time it takes to restore the data one may be better off running down to the local drug store and grabbing one of the 1 or 1.6 TB boxes on sale these days.

The talk about parking drive heads sends chills up my back. I remember worrying about this 10-15 years ago but hadn't heard boo about it since. But this is my first venture into Linux world so maybe it is common. But the same box and drives I am having a problem with ran Win95 and could withstand plug pulls with no problem for years.

I think the NASlite is great, I have the UPS protecting it and the network so it shuts down fine with the script launched from another machine and if not I can always read the drives with XP. But it would be nice if there was some information on what systems have this power down issue and which ones have no problems.

On a controlled shut down something is written to the HDDs (I see the drive light working ) so it is a system issue, in my mind.

cheers

Speculating about what is or may be happening, without any facts can only create more confusion about an otherwise common condition. I’ll attempt to explain what you are experiencing with the hope to shed some light on the issue.

First, one must understand that IDE is a processor-driven interface. If you pull the plug, the processor doesn’t just stop what it’s doing at the completion of an instruction but it crashes, often right in the middle of something. Although that is not always a problem, it may cause anomalies in the currently mounted filesystems.

Next item to consider is that as most operating systems, NASLite (Linux) buffers the data to be input or output to the physical disk. Buffering is necessary in order to avoid accessing the drives for each and every byte. Without buffering, disk access will be very inefficient and extremely slow.

If you have buffered data and the processor stops in the middle of things, filesystem anomalies are sure thing. Knowing that it is easy to see why NASLite is so particular about verifying and fixing the filesystem prior to mounting it for export.

Another thing to consider is that NASLite often pushes hardware to it's limits where IDE interfaces designed for 8G drives are addressing 300G drives. Although that is perfectly acceptable, some IDE hardware may have bugs that may appear and cause problems with some drives or drive types.


The “parking drive heads” post was simply a discussion about drives that use disk inertia to place the heads in the home (parked) position and drives that don’t do that, hence dropping the heads right on top of the data area of a disk. No reason for chills , just some factual information.

Lastly, the reason one may see the drive LEDs flicker when NASLite is properly shut down: services are killed, buffers are flushed, and drives are unmounted. Flushing the buffers in particular since that step will write any buffered data to the drives prior to unmounting.

I hope the above makes sense. To avoid filesystem problems, my advice is to be mindful of the hardware issues and allow NASLite to complete the filesystem verification on boot. It is not there to be a pain but to ensure the integrity of your data

If there are any other questions regarding this, I’ll be happy to provide answers where I can.

I too now have a problem after upgrading NASLite to v1.5. (USB)

After successfully reconfiguring the OS (with HDs disconnected, as I always do to avoid surprises), I rebooted with HDs connected.

Boot hangs at the "verifying dmi pool data...", although it will boot if drives are disconnected, and actually did boot once with one HD connected, although it wasn't detected.

I remember turning off the box once without shutting it down first, and having to wait for the OS to check all drives (quite long, I have 850MB in total), but I think I was able to telnet to it (as I did to check all was ok).

Refer to http://www.serverelements.com/phpBB2/viewtopic.php?t=867