Secure Erase SSD outside Windows?

[Hopper642]

I will probably reinstall windows 7 fairly soon, and I have 2 Intel 520 SSDs in RAID 0. I was going to delete the current RAID array where the OS is installed and then do a secure erase outside windows. How would I go about this? I have a boot disc DVD of Acronis, and I have used the disc erase feature with this software, but is this the best way to do this, or should I use Intel's SSD toolbox in windows for this task? I do have another machine where this could easily be done (that is, I can plug these two drives into the second PC). Thanks.

 

[Coup27]

You have the following methods available in order of ease:

1. Connect the drive as a storage drive on another system and use Intels SSD Toolbox.

2. Use this. This is the official secure erase ISO utility. You can boot this and issue a secure erase command. I've put this after the Intel method as I've never personally got this program to work on any PC I've tried it on.

3. Use this. It's a lot more of a faff but it does definately work.


This thread highlights how criminal it is of Intel not to include their own bootable secure erase utility like Samsung do. I can secure erase my 830 with their bootable utility in no time.

 

[Hopper642]

Thanks for the information. Very helpful.

 

[razel]

If you are just reinstalling Windows 7, then there is absolutely no need for a secure erase. It sounds like you want to remove the RAID setup to go back to a single drive. If so once you get that single drive going with Win7 installed, then you can use Intel's Toolbox to secure erase the remaining drive. But I'd only do that if I plan on selling it.

 

[Coup27]

True there is no need if you are just reinstalling Windows, but reading another thread he posted in he did experience an error of some description so SE is a good way to totally clear the drive before starting again. For how long it takes I think it's worth doing IMO.

 

[Hopper642]

That's exactly correct. The rationale for the SE is the fact that I had an error with this RAID array initially. Windows kept wanting a scan disc. No errors were found. This happened to me about 3-4 times. I then saw a burb about reenabling indexing service. I had disabled indexing as was recommended for SSDs. After reenabling this, I have had no issues with windows 7, but Acronis still sees a problem when I try to do a backup. I don't know if it is one of the drives or Acronis. Regardless, I was going to update my video cards and do a SE at the same time and then reinstall windows. I was planning to keep the RAID array too. I just reinstall windows every few months anyway. So much junk gets in the registry that I have found that my stuff functions much better with a fresh install of the OS (drivers and the like). I am sure opinions vary on this, but it is just what I have observed. Thanks for the input from you both. Much appreciated and helpful.

 

[nine9s]

I can secure erase my 830 with their bootable utility in no time.

I am getting a 830. How does the secure erase work?

 

[Coup27]

You have two methods. Either you can SE a storage drive via the Magician software in the same manner the Intel SSD Toolbox does, or you can create a bootable CD or USB stick from within Magician.

Once you boot this you follow the steps and perform the SE. Very easy and takes less than a minute.

 

[groberts101]

I will probably reinstall windows 7 fairly soon, and I have 2 Intel 520 SSDs in RAID 0. I was going to delete the current RAID array where the OS is installed and then do a secure erase outside windows. How would I go about this? I have a boot disc DVD of Acronis, and I have used the disc erase feature with this software, but is this the best way to do this, or should I use Intel's SSD toolbox in windows for this task? I do have another machine where this could easily be done (that is, I can plug these two drives into the second PC). Thanks.

I'll butt in,and then right back out, to hint you this.

The little known fact is that you do not necessarily use a utility to secure erase the SSD. It's actually more like.. the utility you use MUST access and communicate with the drives INTERNAL ATA SE protocol which is inbuilt on all HDD/SSD firmware as the federal law requires. HDD get's 00's written(which is why it takes so long).. and Sandforce SSD gets flooded with a sudden voltage charge to physically set the nand back to a logically empty state(all flashed back to the SSD required 11's in seconds).

So, with that in mind, the "disk erase" feature in Acronis is like the older low level format in which data is actually written to disk in attempt overwrite(since data in never physically "erased") all logical space. Now, with a controller like the SF-228x?.. you get additional data written to flash(because Acronis doesn't detect SSD and assumes it is HDD). AND then.. you get a dirtied up and diminished fresh block reserve as the drive maps that new data and tries to make sense of what the hell you just wrote to it. lol

Even trim will not save you with that particular controller and recovery will often be deferred to lower activity recovery times on Sandforce. That's when garbage collection kicks in and does most of the work.

IMHO?.. never.. ever.. reinstall to a Sandforce based drive without first secure erasing it to properly rebaseline to clean factory state. I've helped hundreds and hundreds of folks who ran into all sorts of mapping/speed issues with these types of SF drives and the bulk of them were remedied with secure erase/reinstall protocol's. CMOS clears are also a very important aspect of fresh baselines as well.

I even go so far as to reflash the firmware during the SE process. The very few destructive flashes that are floating around the net can be even better as they reset the entire drive right down to the basecode which pretty much gives you an "as fresh as you'll ever get" scenario.

Revisional firmware can change algorithms and pump up some area's of the controllers efficiency(often at the loss of another area because the "processing power pie" can get no larger with only the size of the alotted pieces changing).. BUT it cannot change the inherent inner limitations of the controllers processor. That amount of code work is usually called.. "a new controller". lol

Serously though, even Intel is not accessing Sandforce's basecode IP and all they can do is the same thing the other guys are doing with their firmware. They test it.. it breaks.. they tweak it. They just have more money to test it.. so you don't have to break it and then help teach them what's going on with it.

Not debating.. just sharing what I know from beta-testing Sandforce drives and measuring the inherent limitations that exist with them no matter the mfgrs using that chip in their drives.

Sorry I rambled a bit more than usual there. Got my multitasking time split up and kept adding to it over about an hour. Hopefully you got the main points mixed in there. Good luck with it all.

 

[Hopper642]

I'll butt in,and then right back out, to hint you this.

The little known fact is that you do not necessarily use a utility to secure erase the SSD. It's actually more like.. the utility you use MUST access and communicate with the drives INTERNAL ATA SE protocol which is inbuilt on all HDD/SSD firmware as the federal law requires. HDD get's 00's written(which is why it takes so long).. and Sandforce SSD gets flooded with a sudden voltage charge to physically set the nand back to a logically empty state(all flashed back to the SSD required 11's in seconds).

So, with that in mind, the "disk erase" feature in Acronis is like the older low level format in which data is actually written to disk in attempt overwrite(since data in never physically "erased") all logical space. Now, with a controller like the SF-228x?.. you get additional data written to flash(because Acronis doesn't detect SSD and assumes it is HDD). AND then.. you get a dirtied up and diminished fresh block reserve as the drive maps that new data and tries to make sense of what the hell you just wrote to it. lol

Even trim will not save you with that particular controller and recovery will often be deferred to lower activity recovery times on Sandforce. That's when garbage collection kicks in and does most of the work.

IMHO?.. never.. ever.. reinstall to a Sandforce based drive without first secure erasing it to properly rebaseline to clean factory state. I've helped hundreds and hundreds of folks who ran into all sorts of mapping/speed issues with these types of SF drives and the bulk of them were remedied with secure erase/reinstall protocol's. CMOS clears are also a very important aspect of fresh baselines as well.

I even go so far as to reflash the firmware during the SE process. The very few destructive flashes that are floating around the net can be even better as they reset the entire drive right down to the basecode which pretty much gives you an "as fresh as you'll ever get" scenario.

Revisional firmware can change algorithms and pump up some area's of the controllers efficiency(often at the loss of another area because the "processing power pie" can get no larger with only the size of the alotted pieces changing).. BUT it cannot change the inherent inner limitations of the controllers processor. That amount of code work is usually called.. "a new controller". lol

Serously though, even Intel is not accessing Sandforce's basecode IP and all they can do is the same thing the other guys are doing with their firmware. They test it.. it breaks.. they tweak it. They just have more money to test it.. so you don't have to break it and then help teach them what's going on with it.

Not debating.. just sharing what I know from beta-testing Sandforce drives and measuring the inherent limitations that exist with them no matter the mfgrs using that chip in their drives.

Sorry I rambled a bit more than usual there. Got my multitasking time split up and kept adding to it over about an hour. Hopefully you got the main points mixed in there. Good luck with it all.

Don't apologize at all. That's the clearest, most concise explanation of this technology that I have run across yet. I did wonder why a SE with Intel's SSD toolbox was so quick, and you explained that quite well. The longer I have this machine and these parts, the more I pick up here and there. I am in your debt. I will undoubtedly move these drives over to a second machine to do a SE via Intel's own toolbox assuring me of the proper result as you describe.

I could even reflash the firmware at the same time, but how critical is this overall? And, also, how critical is "CMOS clears"? I am assuming you are not speaking of the CMOS with the motherboard, correct?

Thanks for the help. Most appreciated.

 

[taltamir]

I will probably reinstall windows 7 fairly soon, and I have 2 Intel 520 SSDs in RAID 0. I was going to delete the current RAID array where the OS is installed and then do a secure erase outside windows.

Why do you need to secure erase the drive if you are going to reinstall windows on it?

 

[VirtualLarry]

The little known fact is that you do not necessarily use a utility to secure erase the SSD. It's actually more like.. the utility you use MUST access and communicate with the drives INTERNAL ATA SE protocol which is inbuilt on all HDD/SSD firmware as the federal law requires.

I have never heard that support for ATA "Secure Erase" is in any way mandated by law. In fact, I highly doubt it is, because if it was, then things like USB flash drives would likewise be covered, and the fact is, they entirely lack any type of "Secure Erase" functionality.

In short, please provide a reference, because I think that you are incorrect.

 

[lsv]

I have never heard that support for ATA "Secure Erase" is in any way mandated by law. In fact, I highly doubt it is, because if it was, then things like USB flash drives would likewise be covered, and the fact is, they entirely lack any type of "Secure Erase" functionality.

In short, please provide a reference, because I think that you are incorrect.

I'd love to have a tool like this, the last thing I want is to write to the entire drive for no reason when selling an old SSD.

Hehe 'old'.

 

[Shmee]

also consider GParted for that. I have had to secure erase my 2nd agility 2 twice now to get it working again. Gparted made it much easier, a very good program.

 

[Hopper642]

Why do you need to secure erase the drive if you are going to reinstall windows on it?

I don't necessarily think I have to do this, but with a couple of error messages with these drives in windows, I thought it would be a good idea to make sure whatever is going on is completely wiped out. I thought the SE feature with the Intel SSD Toolbox would be the best way to go.

 

[VirtualLarry]

http://mackonsti.wordpress.com/2011/11/22/ssd-secure-erase-ata-command/

Instructions for doing a SE under Linux using hdparm.

I've used these instructions myself, they seem to work.

 

[taltamir]

I don't necessarily think I have to do this, but with a couple of error messages with these drives in windows, I thought it would be a good idea to make sure whatever is going on is completely wiped out. I thought the SE feature with the Intel SSD Toolbox would be the best way to go.

What error messages?
This doesn't sound like something that secure erase will fix, so lets start diagnosing your problem from the beginning and see if we can help you fix your system.

 

[razel]

I have never heard that support for ATA "Secure Erase" is in any way mandated by law. In fact, I highly doubt it is, because if it was, then things like USB flash drives would likewise be covered, and the fact is, they entirely lack any type of "Secure Erase" functionality.

In short, please provide a reference, because I think that you are incorrect.

From what I remember reading from the UC San Diego group. It was not a law, but a requirement for hard disks purchased by some government departments. I myself am surprized at how widely supported SE is. I have a 2000 Hitachi 30GB 4200 RPM notebook drive and even it supported 'enhanced secure erase.'

 

[Hopper642]

What error messages?
This doesn't sound like something that secure erase will fix, so lets start diagnosing your problem from the beginning and see if we can help you fix your system.

Thanks. What was happening occurred on boot up of the machine. This was before it was actually in windows. I received a message after the bios post and before the windows desktop that windows wanted to check the disc for errors. I allowed it to check for errors, but nothing was found. Acronis found an unreadable section before it made a backup (outside the windows environment with a boot disc), and I simply ignored this and allowed the backup to proceed. This happened about 3-4 times. I then saw an article on file indexing. It was mentioned that reenabling file indexing may solve this issue, and it did (or I thought it did). I had disabled indexing as was recommended for an SSD when I originally installed the RAID array.

The specific error? As I check the event viewer as I am typing this-


Log Name: System
Source: Ntfs
Date: 4/16/2012 7:32:09 PM
Event ID: 55
Task Category: (2)
Level: Error
Keywords: Classic
User: N/A
Computer:
Description:
The file system structure on the disk is corrupt and unusable. Please run the chkdsk utility on the volume \Device\HarddiskVolume2.
Event Xml:
<Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
<System>
<Provider Name="Ntfs" />
<EventID Qualifiers="49156">55</EventID>
<Level>2</Level>
<Task>2</Task>
<Keywords>0x80000000000000</Keywords>
<TimeCreated SystemTime="2012-04-17T00:32:09.612638700Z" />
<EventRecordID>9885</EventRecordID>
<Channel>System</Channel>
<Computer>Hopper64-PC</Computer>
<Security />
</System>
<EventData>
<Data>
</Data>
<Data>\Device\HarddiskVolume2</Data>
<Binary>00000C000200380002000000370004C000000000320000C0180000000000000000000000000000003F200A00FFFFFFFFFFFFFFFF</Binary>
</EventData>
</Event>


Now as I look at the event viewer, I have one today, the first since 4/16/12:

Log Name: System
Source: Ntfs
Date: 5/12/2012 6:05:51 AM
Event ID: 55
Task Category: (2)
Level: Error
Keywords: Classic
User: N/A
Computer:
Description:
The file system structure on the disk is corrupt and unusable. Please run the chkdsk utility on the volume \\?\Volume{e2450b7b-80e4-11e1-9465-806e6f6e6963}.
Event Xml:
<Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
<System>
<Provider Name="Ntfs" />
<EventID Qualifiers="49156">55</EventID>
<Level>2</Level>
<Task>2</Task>
<Keywords>0x80000000000000</Keywords>
<TimeCreated SystemTime="2012-05-12T11:05:51.265617100Z" />
<EventRecordID>17012</EventRecordID>
<Channel>System</Channel>
<Computer>Hopper64-PC</Computer>
<Security />
</System>
<EventData>
<Data>
</Data>
<Data>\\?\Volume{e2450b7b-80e4-11e1-9465-806e6f6e6963}</Data>
<Binary>0D000C000200380002000000370004C000000000320000C00000000000000000000000000000000080091200FFFFFFFFFFFFFFFF</Binary>
</EventData>
</Event>


But I also have a warning that this was repaired today too:

Log Name: System
Source: Ntfs
Date: 5/12/2012 6:06:13 AM
Event ID: 130
Task Category: None
Level: Warning
Keywords:
User: SYSTEM
Computer:
Description:
The file system structure on volume C: has now been repaired.
Event Xml:
<Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
<System>
<Provider Name="Ntfs" Guid="{DD70BC80-EF44-421B-8AC3-CD31DA613A4E}" />
<EventID>130</EventID>
<Version>0</Version>
<Level>3</Level>
<Task>0</Task>
<Opcode>0</Opcode>
<Keywords>0x8000000000000000</Keywords>
<TimeCreated SystemTime="2012-05-12T11:06:13.199255600Z" />
<EventRecordID>17026</EventRecordID>
<Correlation />
<Execution ProcessID="4" ThreadID="700" />
<Channel>System</Channel>
<Computer>Hopper64-PC</Computer>
<Security UserID="S-1-5-18" />
</System>
<EventData>
<Data Name="VolumeIdLength">2</Data>
<Data Name="VolumeId">C:</Data>
<Data Name="RepairDetail">25008: Start repair on 05/12/2012 at 06:06:13:199
26084: Deleting an index entry from index 0x1000000000c53 of file 0x2a0000000004d3.
27094: An index entry of index 0x1000000000c53 points to unused file 0x2a0000000004d3.
27293: File record 0x2a0000000004d3 maps to "\Windows\System32\LogFiles\WMI\RtBackup\EtwRTEventlog-Security.etl".
25009: End repair on 05/12/2012 at 06:06:13:199
</Data>
<Data Name="RepairDataLength">140</Data>
<Data Name="RepairData">D304000000002A00BB0117000600000000000000000000000000000000000000000000000000000000000000000000000000000000000000010000000000000000000000000000000000000000000000000000000100000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000</Data>
</EventData>
</Event>


Thanks for your help.

 

[WhoBeDaPlaya]

Simplest way :

1) Create gparted bootable disc
2) Disconnect all other HDDs
3) Boot up gparted distro, close gparted and open terminal
4) "hdparm --security-erase NULL /dev/sda"
5) DONE

I use this all the time when reimaging my OS.

 

[taltamir]

This was before it was actually in windows. I received a message after the bios post and before the windows desktop that windows wanted to check the disc for errors. I allowed it to check for errors, but nothing was found.

1. Disabling indexing is nonsense for SSDs. It was suggested because first gen jmicron SSDs had worse random write performance then HDD (modern SSDs are 100x faster on random writes then HDD).
2. Windows automatically checks the disk for errors when the system was shut down improperly while the drive was working, or a new drive was added. Neither of them indicate that there is ANYTHING wrong with your drive.
3. Secure erase will do nothing for you.

So, what we need to figure out is why windows is doing it. I am thinking you should try swapping the SATA cable (maybe it has a connection issue and gets disconnected/reconnected). Aside from that, its just a matter of checking the other hardware...

Question, do you ever power down the system by just cutting power via power button (it is possible to configure in bios for the power button to not instantly cut power and use it to do a proper shutdown by telling it to wait 4 seconds and then pressing for less than 4 seconds; basically I Am asking, do you hold or just press quickly the power button and the PC instantly turns off without going through the "windows shutting down" process and screens)

 

[Hopper642]

1. Disabling indexing is nonsense for SSDs. It was suggested because first gen jmicron SSDs had worse random write performance then HDD (modern SSDs are 100x faster on random writes then HDD).
2. Windows automatically checks the disk for errors when the system was shut down improperly while the drive was working, or a new drive was added. Neither of them indicate that there is ANYTHING wrong with your drive.
3. Secure erase will do nothing for you.

So, what we need to figure out is why windows is doing it. I am thinking you should try swapping the SATA cable (maybe it has a connection issue and gets disconnected/reconnected). Aside from that, its just a matter of checking the other hardware...

Question, do you ever power down the system by just cutting power via power button (it is possible to configure in bios for the power button to not instantly cut power and use it to do a proper shutdown by telling it to wait 4 seconds and then pressing for less than 4 seconds; basically I Am asking, do you hold or just press quickly the power button and the PC instantly turns off without going through the "windows shutting down" process and screens)

I think that I may have hit ctrl-alt-del once before I had installed windows. That, of course, did not power the system off completely, and windows was not installed at that point. I have never cut the power to this system without windows shutting down completely and properly.

I will try to change the sata cables, but I really thought there was just a simple corrupt file somewhere. Wiping the whole thing seemed reasonable. The error involves the file structure in the OS, so it seemed logical to purge the whole thing when I reinstalled windows when I upgraded my video card. I am just doing this due to prior driver changes and the junk that gets left over in the registry.

Thanks for the help. Appreciate your time and efforts.

 

[Hopper642]

BTW-Once I reenabled indexing, I never had problems again with a forced chkdsk on bootup. I still see ntfs errors occasionally as above.

 

[taltamir]

I think there is something wrong somewhere in your system. PSU, RAM, CPU... something is causing an issue. And I think the indexing thing was coincidence and it reoccur.

 

[Hopper642]

I think there is something wrong somewhere in your system. PSU, RAM, CPU... something is causing an issue. And I think the indexing thing was coincidence and it reoccur.

Certainly possible, but I don't think so only because this was not happening until I installed the 2 Intel SSDs. I previously had a single SSD running this system with no problem until I went to the 2 520 Intel SSDs in RAID 0. I never had a forced chkdsk at boot until I installed the 2 new SSDs in RAID.

It would seem more logical that whatever is causing this has everything to do with that change. It would be a great coincidence otherwise since I had been using the same system with the same components for months without a ntfs error or forced chkdsk at boot. ???

Thanks again for your help. I still think there's a corrupt file in there somewhere that's causing the issue.