NASLite Network Attached Storage

Hi,

I can't see the use of the file system check every x mounts. I understand the need for file checks to keep the system clean but the system used to start them seems not the best.

I have 3 disks and stop the server every night which results for me waiting the file check 3 times every month... OTOH, there are people on this forum that almost never stop the server and they will not have any file check for several years!

I think that it would be much more logical if the file system check was started every x hours of use or every x GB written / read.

Just my 2 cents for version 2.07 (or 3.0 ? )

Regards

It does seem rather odd that there is only a check run when the box is rebooted, might be nice if you could initiate it from the console or web admin. I leave mine on 24/7 so your right about not having a file check run very often.

As an aside, why would you go and shut down your NAS box every night? You do realize that you are causing ware and tare on the drives and the MB/processor combo, Right? The thermal cycling is hard on the parts in the computer and can lead to premature failure. Also, every time you spin down a drive the heads contact the parking surface of the platter and ware down a bit. While the platters are spinning there is no contact between them as the head floats on a cushion on air drug into the gap, hence no ware. The down side is that the spindle bearings are waring but these seem to fail far less than do the drive electronics.

I leave my box on for the above reasons and also because it's a good way to gauge the stability of the particular combo of board, processor, memory, NIC, and NL. I have a board/processor combo that did not work well under NL and will lock under load. A change of MB/Processor and not a problem after moving over 1/2 terabyte.

Mike

Hi Mike,

The reason is to use less energy and try to leave a cleaner world to my kids Now, this said, I had not taken into account the energy needed to rebuild the parts that may die faster... I'll need a new computer and a special program just to see if it is better to turn off the server and the 3 computers or leave them on

Regards.

I can understand your reason and you went just where I was going to go IE the cost of making the new parts to replace those that fail. Just keep in mind that NL needs little in the way of horse power to get it's job done and so a low power/ low speed processor will be more than enough. Though I can't tell you if they are stable or not there are a few low power offerings from a couple of companies, some of which don't even require a heat sink with a fan. Also the HDD shouldn't eat much in the way of power if it is just idle, they draw far more power spinning up and during read/write. I wouldn't be surprised if you couldn't build a NAS box that only burnt 20-25 Watts idle and maybe 35 full tilt.

Mike

mikeiver1,

The filesystem check done during power-up is a check and repair which requires the filesystem to be RO in order to avoid problems with open or locked files. Obviously, the best time to do this is during boot. The only thing it doesn't do is bad block testing since that is SOOOO time consuming.

You can perform a RO filesystem check while the system is running using the admin menu. That test will tell you if there are problems with the filesystem, but no changes are written to the drive. A complete check and repair is usually not required unless you suspect problems with your drive or if the check and repair at boot is unable to complete the task. If you find that the filesystem test reports a problem you can reboot to allow NASLite-2 to check and repair at boot. Rebooting should not spin the drives down and will result in a quickest recovery. To do a complete check and repair with scan for bad blocks, you have to set the suspect drive to idle ( -- ) so it is not mounted or shared, then reboot and use the admin to do the check and repair. When done, reset the drive back to RO or RW and reboot to remount.

I think that approach is very robust and failsafe. It protects novices from themselves while allowing sensible admins to do a more agressive storage upkeep.