making use of my 2nd nic in my unRAID server

one last thing:

 

i've shut down my unraid server because i'll be doing some surgery on it asap (adding two more sata drives, swapping sata optical for ide(*), etc.) and discovered that i'm apparently missing a mail program (what is listed in the 'apcupsd' dependencies)... and a 'zip' one to shrink log files.  i've looked in the unmenu package manager page to no avail.

 

help?

As far as I remember, "infozip (Info-ZIP's zip and unzip utilities)" is one of the packages in the package manager page.

 

As far as mail, it is not needed, but you can add it if you like.  There are many mail scripts posted.

 

One combination mail and notify script is here: http://lime-technology.com/forum/index.php?topic=2470.0

...and another:

 

i know the parity drive must be as big as the biggest data drive, but is there some rule of thumb saying "beyond this amount of data, your parity drive should be (size of biggest drive) + xx%"?

 

i am about to have 6 drives in my unraid server (2x1 terabyte data + 3x500 megabyte data + 1 terabyte parity) and am starting to wonder if the parity drive will be enough to do its magic.

You do not understand how parity works.  A single parity drive can hold the correct party calculation for an infinite number of data disks of equal or smaller size.

 

Each bit on the parity drive is calculated based on the values of the equivalent bit on all the other data drives.  In unRAID, if there is no equivalent bit, because a particular data drive is smaller than the parity drive, a zero is used for the bit value for that drive.  Basically, all the bits on a given bit position when added to the bit value stored on the parity drive must equal an even number.      the parity bit is set to whatever is needed to end up with an even number.  If all the other drives have zeros as their data values in that bit position, the parity drive must also contain a zero for an "even" number.  If two of the data disks had "ones", the parity drive must also have a zero, again for an even number (in this case "2")  It does not matter which disks have the ones, and which have the zeros, all we care about is if it is an odd number, or an even, so the server can set the parity bit to make the total number across a given bit position an even number of bits set to "1".

 

The magic comes when one disk fails.  If we have a parity calculation, and we can read all the remaining disks for their bits in a given bit position, we can deduce the missing disk since we know the total number of bits set MUST be even.  If we read the remaining disks and find an odd number of bits set to a "1" then the missing disk's bit must have also been a "1" to end up with an even number.  repeat this deductive logic trillions of times for a terabyte drive and you can completely reconstruct what it was when it could be read.

 

Now, to calculate parity, or reconstruct a missing/failed disk all the other disks in the array must be read to get their equivalent sectors to reconstruct the sector on the disk that cannot be read.  The bus-throughput limits the number of disks that can be read and still supply data to play a movie, etc. (Hardware limits how many disks can be involved too).    The parity calculation is repeated trillions of times for a terabyte drive, and reading every bit of the entire set of drives to initially calculate parity or later to check it takes many hours for large arrays.

 

See here: http://lime-technology.com/forum/index.php?topic=2005.msg14572#msg14572