Jump to content

Pre-clear versus Parity - Newbie trying to understand unRaid


ChuckS

Recommended Posts

I am new to unRaid and trying to understand some of the basic concepts. I have built a system following the guides here on the forms and WiKi. I was a bit confused on the pre-clear script.  It seemed in the user guide that it mentioned it was not necessary to run a pre-clear when setting up a system for the first time.  It also mentioned to copy all your data then enable the parity drive as this would save some time. I started my initial install without pre-clearing the drives and with parity off (I reserved a drive but am not using it yet.)

 

So my question is around parity and pre-clear. It seems one reason to run the pre-clear script when adding a new drive to an already parity protected system is that the parity would not be effected by a drive with all zeros. So if this is true and parity is not used then there should be no reason to run pre-clear (other than testing the drive for potential issues). If parity is turned on then why couldn’t a new drive (without being cleared to all zeros) be added if you run a new parity check right after adding? Seems like this might be faster then pre-clearing? Not sure what I am missing or if my initial install was not done correctly (although it seems to be working for the most part.)

 

Thanks,

 

Chuck

 

Link to comment

One in five new drives will fail the pre-clear. Those failing drive will fail shortly if no pre-clear is done. The odds are against your build being successful. You should pre-clear all new drives even those used in other systems.

 

Adding data without a parity disk in place means that you have no idea whether the data can be read back. This is asking for trouble.

 

UnRAID will not allow a drive that has not be clear be added to an array with parity in place. A parity check is not the same as generating parity. Generating parity will take as long as a clear.

Link to comment

I am new to unRaid and trying to understand some of the basic concepts. I have built a system following the guides here on the forms and WiKi. I was a bit confused on the pre-clear script.  It seemed in the user guide that it mentioned it was not necessary to run a pre-clear when setting up a system for the first time.  It also mentioned to copy all your data then enable the parity drive as this would save some time. I started my initial install without pre-clearing the drives and with parity off (I reserved a drive but am not using it yet.)

 

So my question is around parity and pre-clear. It seems one reason to run the pre-clear script when adding a new drive to an already parity protected system is that the parity would not be effected by a drive with all zeros. So if this is true and parity is not used then there should be no reason to run pre-clear (other than testing the drive for potential issues). If parity is turned on then why couldn’t a new drive (without being cleared to all zeros) be added if you run a new parity check right after adding? Seems like this might be faster then pre-clearing? Not sure what I am missing or if my initial install was not done correctly (although it seems to be working for the most part.)

 

Thanks,

 

Chuck

 

new drives are not guaranteed to be all zeros.  You cannot ever make that assumption.

 

 

Link to comment

Thanks for the responses. I am having some other NIC related (I hope) issues so I will probably redo the build and pre-clear the drives. So far only a test system so nothing really lost yet. Would rather do it right.

 

On the fact that 1 in 5 drives fails pre-clear what do others do about this? Seems like a very high faliure rate. Do you return the drives to the manufacturer? Seems like a 20% cost increase on drives if you just eat that percentage of bad drives. Any issue with manufacturers taking them back?

 

Regards,

 

Chuck

Link to comment

Writing all 0's to the whole data disk or writing new parity data to the whole parity disk will both take the same amount of time. Writing 0's will allow the array to remain protected and re-writing parity will leave the array unprotected. OK, the data disk should be quicker if it was smaller than the parity disk.

 

unRAID can write 0's by itself. It will take about 8 hours for a 2T disk and your array will be offline during this time. However, you will have no idea if the drive is 100% healthy after this is done. You need to run a parity check once the clearing is done to know everything is OK. So, 8hrs for the clearing and 8 hours for the parity check vs running the preclear with no array down-time and then adding it which may take about 10 minutes to add and format. I hope now it's clearer why the preclear is the way to go.

 

You can also use switches to bypass the pre-read and post-read passes of the preclear script so that it takes the same time as unRAID. I have done this before on seasoned drives to add them to the array quicker. Details should be in the preclear script thread or in the preclear script help (type preclear.sh /h or something like that to get a help page). I would never do this on a new drive though.

 

The preclear script really is pretty amazing work and a number of people here have adapted it as their standard burn-in test for every single drive they purchase no matter the final use for that drive.

Peter

 

Link to comment

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...