[Solved] Advanced Format Question

[TODDLT]

I am in the process of building my first unRAID server and just started running pre-clear on 4 drives simultaneously.

 

I have:

2 - Samsung HD204UI

1 - WD 20EARS

1 - Hitachi HDS7222020ALA330.

 

I thought the two samsung's and the WD were Advanced Format drives, and the Hitachi specifically says on the specs that it is 512K / sector.  So I started the 3 with -A switch and the Hitachi normally.  When I was starting the 2nd Samsung I noticed somewhere it said 512K sector.  So can anyone confirm which ones are Advanced format and which ones aren't?  If I have it wrong, How do you terminate the preclear to start over.  They have only been running a few minutes.  I did install the screen package so just dropping the telnet connection shouldn't stop the preclear.

[S80_UK]

Preclear with -A will not have any adverse effect on normal drives with 512-byte sectors [not 512K, BTW]. 

 

To explain it a little (although it's well covered elsewhere) the -A parameter does not affect the size of the sectors used on the drive, it just varies the position of the first used sector by one 512-byte sector location so that it is aligned to a multiple of 8 512-byte sectors (specifically to sector 64 instead of 63).  In this way the first used sector is also aligned with the sector boundary on any drives that use the advanced format (they use 4096-byte sectors currently). 

 

This means that you do not have any need to abandon the preclear operations that are currently in progress.

[TODDLT]

Preclear with -A will not have any adverse effect on normal drives with 512-byte sectors [not 512K, BTW].  

 

To explain it a little (although it's well covered elsewhere) the -A parameter does not affect the size of the sectors used on the drive, it just varies the position of the first used sector by one 512-byte sector location so that it is aligned to a multiple of 8 512-byte sectors (specifically to sector 64 instead of 63).  In this way the first used sector is also aligned with the sector boundary on any drives that use the advanced format (they use 4096-byte sectors currently).  

 

This means that you do not have any need to abandon the preclear operations that are currently in progress.

 

Thanks for clearing this up and of course restating the obvious at the end .  

 

I did understand the difference starting in sector 64 instead of 63.  However, I was not clear on the impact of doing so on a drive that was not advanced format.  IE if it would slow down the process or result in a bad format.  

 

Is it better to just format them all -A to be sure then?

 

Thanks again for clearing it up.

[S80_UK]

You can do the Preclear on all drives. 

 

Note that for the WD20EARS which has an optional jumper that can also be used to force an alignment, then the jumper should NOT be fitted if you are using Preclear with -A in order to avoid applying the sector alignment twice which would then cause things to become unaligned again.

 

The exception to that rule is that if the WD20EARS drive has already been used with the jumper fitted, then the safe option is generally to leave the jumper in place and to Preclear that drive without the -A (with a lower-case -a would be better).  Once they've been in use with the jumper fitted, the WD drives don't always like being used with the jumper subsequently removed.

 

Also, once you have the alignment preference specified in the unRAID settings, note that Preclear will use that setting if you do not specify -A or -a in the Preclear command line.

 

Hope that helps more than it confuses... 

[TODDLT]

You can do the Preclear on all drives. 

 

Note that for the WD20EARS which has an optional jumper that can also be used to force an alignment, then the jumper should NOT be fitted if you are using Preclear with -A in order to avoid applying the sector alignment twice which would then cause things to become unaligned again.

 

The exception to that rule is that if the WD20EARS drive has already been used with the jumper fitted, then the safe option is generally to leave the jumper in place and to Preclear that drive without the -A (with a lower-case -a would be better).  Once they've been in use with the jumper fitted, the WD drives don't always like being used with the jumper subsequently removed.

 

Also, once you have the alignment preference specified in the unRAID settings, note that Preclear will use that setting if you do not specify -A or -a in the Preclear command line.

 

Hope that helps more than it confuses... 

 

The HD is in the case being precleared right now, so I'm not sure how any jumpers are set, but I'm confident it is in the original factory setup.  If I read the above correctly, then I should be pre-clearing with the -A switch (which I am).  So I should be good.

 

Th WD drive is running slower than the other drives.  10-20% slower than the samsungs, and they are 5% slower than the Hitachi.

The WD is also running a little warmer than the Samsung's about equal with the Hitachi.  You would expect the Hitachi to be warmer being a 7200 RPM drive, but the WD should have been cooler.  All are fan cooled.

[S80_UK]

In my experience you will always get some temperature difference that is unaccounted for even with identical drives (often a subtle difference in airflow) and you will also get a spread of speeds covering a range of 10-20% especially if doing multiple preclears at the same time (also with identical drives).  I guess it's partly down to how the drives run interacting with how the operating system manages the different tasks.  Personally, I wouldn't be worried unless there are see much bigger differences.

[TODDLT]

In my experience you will always get some temperature difference that is unaccounted for even with identical drives (often a subtle difference in airflow) and you will also get a spread of speeds covering a range of 10-20% especially if doing multiple preclears at the same time (also with identical drives).  I guess it's partly down to how the drives run interacting with how the operating system manages the different tasks.  Personally, I wouldn't be worried unless there are see much bigger differences.

 

I dont see anything that is worrying me, but I think the statistics are interesting.

 

Hitachi HDS7222020ALA330 - running at 100 - 118 MB/Sec (occasionally drops below 100 for a short duration).  35C consistently.  It has stayed roughly 10% ahead of the Samsung drives.

Samsung HD204UI - 2 of these and the stats have stayed virtually identical the whole time.  % complete has stayed identical.  Speeds from 95 to 110 MB/Sec, often matching the Hitachi.  Drives running 31/32C.

WD20EARS - running 85-90 MB/Sec and 35C.  Progress has stayed behind the Samsungs, by 20% or more.  The specs aren't bad or a problem, but it stands out against the other drives.