Migrating into new case and changing/adding drives

[shazza6887]

Hi All,

 

I have Currently been running my 6 x 8 TB drive build with one parity drive for about 4 years. The current case only has a max storage capacity of 6 drives and is not hot swapable. I am wanting to take all my hardware out of the current case migrate  to a 8 bay Silver stone Case with hot swap capability.

 

I currently use 6 sata connections but with the new case i will be looking to add a either a additional 6 port Sata Card or a 8 Port Sas card although they seem to be hard to find here in Aus. My current port capacity on my mb is maxed out at 6 drives.

 

This is my current mb, ASRock H270M-ITX/ac CPU is a Intel® Pentium® CPU G4560 @ 3.50GHz. Only use server for watching stored movies over network.

 

Will this straight swap work? I want to first transfer the existing hardware into the new case and then add 2 brand new drives, make them my new Parity disk 1 and 2 and then turn the old parity drive into a data drive.

 

Could anyone give me a step by step for the best way to go about doing all this?

 

Thank you all Kindly!

[shazza6887]

No one?

[JorgeB]

On 5/2/2023 at 2:05 PM, shazza6887 said:

Will this straight swap work?

Yes, if no RAID controllers are involved, see here for a list of recommend models.

[shazza6887]

5 minutes ago, JorgeB said:

Yes, if no RAID controllers are involved, see here for a list of recommend models.

 

So this card should work?

 

https://www.ebay.com.au/itm/143980512969?chn=ps&_ul=AU&norover=1&mkevt=1&mkrid=705-139619-5960-0&mkcid=2&mkscid=101&itemid=143980512969&targetid=&device=c&mktype=pla&googleloc=9072067&poi=&campaignid=15791083372&mkgroupid=&rlsatarget=&abcId=9300816&merchantid=112562231&gclid=Cj0KCQjwr82iBhCuARIsAO0EAZyMaA5Z6Nb9MjR2rUka0tIY5Yx8mQetUrNRq9F7Fp29MisFSpogkOUaAlSOEALw_wcB

[JorgeB]

Yes.

[Frank1940]

3 minutes ago, shazza6887 said:

So this card should work?

I would contact the seller about the cable lengths.  If they are 1 meter cables, they are probably too, too long for most cases.  Usually 0.5 meter cables are a much better fit. 

[shazza6887]

29 minutes ago, JorgeB said:

Yes.

 

25 minutes ago, Frank1940 said:

I would contact the seller about the cable lengths.  If they are 1 meter cables, they are probably too, too long for most cases.  Usually 0.5 meter cables are a much better fit. 

So my first step is get this card, get new case, transfer components into the new case and install this card, connect up all the existing drives and they should load up just fine using the same unraid installation?

[Frank1940]

With your use case, that is what should happen.  Make sure your PS has the capacity to handle the load on the +12V buss of the two new drives.  (Personally, I would transfer the hardware from the old case and make sure everything is working before I added the two new drives--- as few changes-- at one time --as possible when doing any type of upgrade.)

[shazza6887]

4 minutes ago, Frank1940 said:

With your use case, that is what should happen.  Make sure your PS has the capacity to handle the load on the +12V buss of the two new drives.  (Personally, I would transfer the hardware from the old case and make sure everything is working before I added the two new drives--- as few changes-- at one time --as possible when doing any type of upgrade.)

Good Point, i have to look into the psu thing. Can not all Psu's handle 8 drives? One of these cases, the backplane only needs 3 sata and 1 molex connector to power all 8 drives!

 

These are the cases i am thinking of changing to,

 

https://www.mwave.com.au/product/silverstone-black-ds380-8-bay-hot-swap-sff-chassis-ab53965

 

https://www.centrecom.com.au/silverstone-cs380-case-storage-series-mid-tower-case-black?gclid=Cj0KCQjwr82iBhCuARIsAO0EAZzl3U_mWSTUHoBxvPyWhFUvdjexkId1XUYyf7sUBIrRFCfBn7XodSsaAm0yEALw_wcB

 

 

[shazza6887]

52 minutes ago, Frank1940 said:

I would contact the seller about the cable lengths.  If they are 1 meter cables, they are probably too, too long for most cases.  Usually 0.5 meter cables are a much better fit. 

https://www.ebay.com.au/itm/334534729350?hash=item4de3d2de86:g:tcYAAOSw4dFi~KFO&amdata=enc%3AAQAIAAAA0M7qyWbTc919hL%2Fx9Atn1W2XOIf5UT0j2BJ6m6F6vQiw%2BJTkAoFNibozhQu78GJe8vGfWBlLDeMvZqytNtMCPMQMXJ2OB%2Byul7LUU756R1UvRlQczIBSh3rXqPxqWPtw5Lw6NvEH4Ev9nw7xmbNOZvgxmnZpVL9oJpMjD35SJWsMyCrJq%2FtM2ilxtwEO%2BkcZpCb3Pac0YlZT%2F1IuYD3nKSxLMBwnIDcq07ff0NWMyFxaNkfRuLtbQmqi8265gnB8i2RAe%2Fbzbfen5aMV0xvMjUU%3D|tkp%3ABk9SR_jFr_H8YQ

 

May have to buy seperate, the 1m will still work though right? might just be harder for cable management?

[Frank1940]

On 5/4/2023 at 10:08 AM, shazza6887 said:

 

May have to buy seperate, the 1m will still work though right? might just be harder for cable management?

Depends on your neatness factor and OCD.  Excess cable lengths can also impede air flow.  IMO, you should use the SATA ports on the MB first, then use the ports on the expansion card.  So you will only need one of those cables. (These cables should be the shielded type and most are.  But I would still try to not to wrap the excess of the active ones into a tight ball to make things look 'neat' to prevent crosstalk between the active ones.) 

 

The one meter ones should work just fine aside from the appearance factor.  It looks like this vendor has been around for ten years and had has a good reviews.  So hopefully the quality of the cables reflect his concern for customer satisfaction.  Plus if you have any problem with cables, they can easily be replaced.

 

On 5/4/2023 at 10:06 AM, shazza6887 said:

Good Point, i have to look into the psu thing. Can not all Psu's handle 8 drives?

 

The spin-up inrush current runs between 2-3 amperes per HD.  For servers, it is recommended that that PSU have a single +12V buss.  In the cheaper PSUs, they often have two +12V busses. One for HD and the other for the GPU.  Remember that any modern PSU will shutdown instantly as soon as it detects that any of the rating limits have been exceeded. 

[shazza6887]

49 minutes ago, Frank1940 said:

Depends on your neatness factor and OCD.  Excess cable lengths can also impede air flow.  IMO, you should use the SATA ports on the MB first, then use the ports on the expansion card.  So you will only need one of those cables. (These cables should be the shielded type and most are.  But I would still try to not to wrap the excess of the active ones into a tight ball to make things look 'neat' to prevent crosstalk between the active ones.) 

 

The one meter ones should work just fine aside from the appearance factor.  It looks like this vendor has been around for ten years and had has a good reviews.  So hopefully the quality of the cables reflect his concern for customer satisfaction.  Plus if you have any problem with cables, they can easily be replaced.

 

 

The spin-up inrush current runs between 2-3 amperes per HD.  For servers, it is recommended that that PSU have a single +12V buss.  In the cheaper PSUs, they often have two +12V busses. One for HD and the other for the GPU.  Remember that any modern PSU will shutdown instantly as soon as it detects that any of the rating limits have been exceeded. 

Thank you for replying! any particular reason to use the sata ports on the MB first? or again just an aesthetic thing?

[Frank1940]

Best reason is that you already know that these ports work satisfactory.  I would use the new ports for the new drives.  (I am one who always wants minimize any problems that introducing new hardware might introduce.)

[shazza6887]

On 5/6/2023 at 10:22 PM, Frank1940 said:

Best reason is that you already know that these ports work satisfactory.  I would use the new ports for the new drives.  (I am one who always wants minimize any problems that introducing new hardware might introduce.)

Hi Frank1940,

 

I am about to build the new Rig, i ended up going with a Fractal Define R5 and also purchased a new PSU - Corsair HX750 Platinum :).

 

Couple of questions, when i install everything before i assign anything i should pre clear both brand new drives?

 

Can you pre clear 2 drives at the same time?

 

Once pre cleared what is the best course for me to take? Assign one of the new drives as parity 2? Then assign current parity 1 drive as either nothing or disk 6, then assign a the other new drive as Parity 1?

[Frank1940]

12 minutes ago, shazza6887 said:

Couple of questions, when i install everything before i assign anything i should pre clear both brand new drives?

It depends on your tolerance to risk.  It is really a hassle if a brand new drive fails in the first few days-- often called 'infant morality'.  (Personally, I always preclear any new drives as soon as I get them.  two reasons, I always want a cold spare on site for instant use in case of a drive failure or the need to add an additional drive for capacity reasons.)

 

28 minutes ago, shazza6887 said:

Can you pre clear 2 drives at the same time?

 

I use the Preclear Docker. below is the link to the FAQ and the PDF instructions on how to use it:

 

      https://forums.unraid.net/topic/83465-support-binhex-preclear/#comment-773206

 

The Docker will allow to clear a fairly large number of drives simultaneously but the speed may well decrease.  ( In the VCN terminal click on File and Add Tab.  You can do one drive in each Tab you have open.)

 

38 minutes ago, shazza6887 said:

Once pre cleared what is the best course for me to take? Assign one of the new drives as parity 2? Then assign current parity 1 drive as either nothing or disk 6, then assign a the other new drive as Parity 1?

 

I would move everything and get the server up and running for a couple of days.  I would then add a new drive as Parity2 drive and build parity2.  Now, if your new drives are both 8TB ones, I would add the second new drive as a data drive.  (If your new drives are bigger than 8TB, then, obviously,  it should be assigned as Parity1.)  If you do end up repurposing the parity drive as a data, I would preclear it first.  That way you should avoid the time required to clear the 'new' data disk and only have to format it.   PS--- When you add  (or replace) any disk, always double check that all SATA connectors (both power and data) to all of the drives are securing seated!!!!

[shazza6887]

On 5/23/2023 at 12:45 AM, Frank1940 said:

It depends on your tolerance to risk.  It is really a hassle if a brand new drive fails in the first few days-- often called 'infant morality'.  (Personally, I always preclear any new drives as soon as I get them.  two reasons, I always want a cold spare on site for instant use in case of a drive failure or the need to add an additional drive for capacity reasons.)

 

 

I use the Preclear Docker. below is the link to the FAQ and the PDF instructions on how to use it:

 

      https://forums.unraid.net/topic/83465-support-binhex-preclear/#comment-773206

 

The Docker will allow to clear a fairly large number of drives simultaneously but the speed may well decrease.  ( In the VCN terminal click on File and Add Tab.  You can do one drive in each Tab you have open.)

 

 

I would move everything and get the server up and running for a couple of days.  I would then add a new drive as Parity2 drive and build parity2.  Now, if your new drives are both 8TB ones, I would add the second new drive as a data drive.  (If your new drives are bigger than 8TB, then, obviously,  it should be assigned as Parity1.)  If you do end up repurposing the parity drive as a data, I would preclear it first.  That way you should avoid the time required to clear the 'new' data disk and only have to format it.   PS--- When you add  (or replace) any disk, always double check that all SATA connectors (both power and data) to all of the drives are securing seated!!!!

Thank Frank!

 

There is the binhex preclear app and also the official pre clear app? Is there any advantage of using the Binhex over the official?

[Frank1940]

The big advantage is that ot is a Docker.  The preclear script is as old as the hills-- think 2010 time period.  Over the years the BASH script has been 'broken' repeatedly by updates to the Linux OS used by Unraid.  This can never happen to the Docker based version as the version of Linux OS that the script uses is a part of the Docker image.  

 

Biggest disadvantage is that you have to use the command line in a Terminal window to preclear a disk.  In that post, I provided is a link to a PDF tutorial which will walk you through the entire set of steps to preclear any disk.  Be sure to read the FAQ also as it contains some tips to proper usage.  You can also copy and paste many of the command line switches required in the Terminal window if you to do it that way. 

Edited May 24, 2023 by Frank1940

[shazza6887]

On 5/24/2023 at 12:10 PM, Frank1940 said:

The big advantage is that ot is a Docker.  The preclear script is as old as the hills-- think 2010 time period.  Over the years the BASH script has been 'broken' repeatedly by updates to the Linux OS used by Unraid.  This can never happen to the Docker based version as the version of Linux OS that the script uses is a part of the Docker image.  

 

Biggest disadvantage is that you have to use the command line in a Terminal window to preclear a disk.  In that post, I provided is a link to a PDF tutorial which will walk you through the entire set of steps to preclear any disk.  Be sure to read the FAQ also as it contains some tips to proper usage.  You can also copy and paste many of the command line switches required in the Terminal window if you to do it that way. 

Ok so far so good, got all the gear in the new tower and array started without issue! 2 new drives popped up in unassigned devices and i am pre clearing one of the new drives now. looks like it will take nearly the whole weekend! I wussed out and just used the old pre clear plugin as i did not want to mess with any terminals etc.

 

So first step, after the pre clear finishes, i stop the array and add the disk as Parity 2, then start the array and let it build it? Once that is done i guess i pre clear the next 10tb and get that ready.

[Frank1940]

40 minutes ago, shazza6887 said:

Ok so far so good, got all the gear in the new tower and array started without issue! 2 new drives popped up in unassigned devices and i am pre clearing one of the new drives now. looks like it will take nearly the whole weekend! I wussed out and just used the old pre clear plugin as i did not want to mess with any terminals etc.

Actually taking that long to do a preclear is good.  The first hundred hours of operation will uncloak virtually all infant morality failures. 

[shazza6887]

On 5/26/2023 at 2:00 AM, Frank1940 said:

Actually taking that long to do a preclear is good.  The first hundred hours of operation will uncloak virtually all infant morality failures. 

Ok so i pre cleared both devices which took approx 39 hours with a perfect result! I have rebuild one drive as Parity 2 and it is now Valid.

 

I am going to now assign the other 10tb drive as parity 1 and let that build... With re assigning the old parity drive as a new data drive do you still recommend doing a full pre clear? Could i just not just 'zero' the drive and add it into the array as it has been used for a long time without issue?

 

Also moving forward would the first data drive i want to replace be this old Parity drive as it would have had the most use out of all the old drives?

 

Cheers!

[itimpi]

4 hours ago, shazza6887 said:

With re assigning the old parity drive as a new data drive do you still recommend doing a full pre clear?

I would just add it (without using preclear) and let Unraid clear it when it is added.   That tends to be the fastest.

 

In terms of which drive it is best to replace that is really up to you.  I would first look through the SMART reports for all drives and see if any are showing signs of potential problems such as pending or reallocated sectors.

[shazza6887]

1 hour ago, itimpi said:

I would just add it (without using preclear) and let Unraid clear it when it is added.   That tends to be the fastest.

 

In terms of which drive it is best to replace that is really up to you.  I would first look through the SMART reports for all drives and see if any are showing signs of potential problems such as pending or reallocated sectors.

Ty itimpi, you guys are so helpful! Is a short smart self test sufficient or should one always run extended tests?

[Frank1940]

39 minutes ago, shazza6887 said:

Ty itimpi, you guys are so helpful! Is a short smart self test sufficient or should one always run extended tests?

Actually, you want to look at the Attributes section/tab for each disk and pay particular attention to the following ones:

 

You may notice that I have left off     'Attribute = 199  UDMA CRC error count'   because it is generally not a Drive issue but (generally) a cabling problem. 

[shazza6887]

9 minutes ago, Frank1940 said:

Actually, you want to look at the Attributes section/tab for each disk and pay particular attention to the following ones:

 

You may notice that I have left off     'Attribute = 199  UDMA CRC error count'   because it is generally not a Drive issue but (generally) a cabling problem. 

Hi Frank, i have just gone through all my disks par the old parity drive as i can see the attributes as it is sitting as an unused disk until i rebuild parity 1,

 

All of these attributes across all my disks are zero except for one disk with regards to attribute '188' which has a raw value of '4295032833'

 

They all have udma errors ranging from 2 to 11

[Frank1940]

OK, I looked at my drives and some did not even report Attribute 188 and others had results greater than zero.   Thinking back, 188 is a special case.  Unraid is reporting a decimal number but the actual value(s) may be a hexadecimal string.  What you have  to do is to google something like this:

 

    seagate  ST6000VN001 attribute 188

 

In my case, Google did give this as a result:

 

    https://superuser.com/questions/1747844/interpreting-seagate-smart-value-188-command-timeout-data-format

 

Hopefully, you will find something similiar.  The problem with SMART is that is not a standard but more of a guideline and the various manufacturers have implemented it differently.  (I believe that Attribute #1 has similiar issues.)  Remember that you are using this as another tool to be used to try to decide which disk to replace next.  Don't assume that that it should be your only decision making tool.