Restoring a messed up SSD
I thought that all of a sudden, someone from the owners of an SSD device would get this topic to think about a backup, someone about a generally more cautious attitude, and would save someone from speaking with a not too rushed support service. Everything written refers not to the devices of the series and the manufacturer that I have.
10 days ago I happened to leave the laptop with the battery in a critical condition overnight without charging. I don’t worry too much about battery life, but the blow came from the other side. In the morning, turning on the laptop in charge and turning it on, I was surprised to find that:
In IOS, the hard drive was determined. Having grabbed the available Ubuntu Live CD on a flash and armed with the command line, I prepared for the debug.
It should immediately be said that in cases of such failures it would be more convenient to use some Data Rescue Live CD, with already installed diagnostic utilities instead of a completely unnecessary office suite, but nonetheless.
We will recruit an arsenal, which will run for us:
We look what happened with us:
So, you can say goodbye to the partition table.
I / O Error? Disk does not support SMART? Already some nonsense.
Yeah. You can see that the number of conditional SSD cylinders dropped 10,000 times and, according to the desktop Gparted, the size of the hard drive is 8MB (I confess, the console command and its output for viewing this outrage are not preserved, please take my word for it). The serial number is missing and instead BAD_CTX is something there. Well, the symptoms understood, you can refer to the search and in support. Indeed, it turns out that the problem is far from single , but, alas, I am the only one with Linux.
In short, for those who are not familiar with the language and lazy, forum users talk about the general exposure of all Intel SSDs to such a bug, especially affecting the 320 series and X25M. There is news about firmware 0362, which is designed to get rid of this particular bug, but the number of people who have already addressed this firmware with the same symptoms indicates an unsolved problem. Yes, the best solution in this case would be to send the hard drive back to Intel, so that they have an incentive to correct their mistakes.
Unfortunately, Intel support is not very efficient, and it responds about once a day, blushing on technical issues, and highly recommending installing their SSD Toolbox to determine the problem. I would like to separately note that the main slice of SSD users is the owners of MacBooks, which, similarly to me, have difficulty installing software under Windows. Special mention is worthy of the fact that this tool, designed to identify faults, requires:
- java
- .NET 3.5
- Windows Media Player Redistributable 11
which makes installing it on a computer loaded with a Live CD an almost impossible task (firstly because of the limitations of the volume on the virtual hard drive, and secondly due to the fact that WMP 11 requires verifying the authenticity of Windows, which, with creaks and groans only Some particularly prominent personalities are obtained in Wine.
Ardent greetings to the developers of this software.
I miraculously managed to explain the situation to the support, and they agreed to a replacement, but to replace, you need to fill in an incredible number of forms, to which you also need to attach confirmation of the device purchased by me. By the will of fate, I am now ten thousand kilometers from home, and did not expect such a trick.
Fortunately, on the forums, everyone unequivocally says that the contents of the disc cannot be restored, but that it is possible to restore the performance. And the time that was spent on correspondence with the support service, I did not spend in vain, but usefully spent reading forums and experiments, a brief summary of which is here and bring.
It is necessary to restore the number of cylinders, returning the cherished 16383.
For this operation, we will need two commands, the launch of both is difficult to protect against a fool and a pest.
We expose the user and password for master operations on the disk.
Next, we need to unlock the extended set of ATA commands, in particular secure-erase, which are blocked when the system boots. There are several ways to do this, one of which is the external box, turn it off and turn it on. I did not have an external box, but miraculously the sending of the laptop to sleep and awakening works.
The following commands perform some kind of safe cleanup, I ran both, since there was no certainty which one would be needed. Before each run the installation of the master password and closing the lid of the laptop.
Now completely for not the faint of heart. Resetting disk settings to factory settings. To launch a command, you need another key that the command line tells you, and to clear my conscience I will not give it here, just as I will mention that this command is marked as SPECIALLY DANGEROUS and DO NOT RUN in the documentation.
Reboot. Hooray, you can put the system.
Total conclusions that I made for myself:
- keep your live cd at hand
- do not leave the laptop completely without power on a critical charge
- make backups, including keyrings, lists of installed packages, configs and rsa keys
- update the firmware (after you know that it works quite well)
- take care of your nerves
I want to additionally note that this method is not always fully restores functionality, and that sometimes the disk remains buggy, braking.
Once again, I send a heartfelt greeting to the support of Intel, and inform them that I still can not log in with my username and password to their community to publish this miraculous recipe from them, and remind you that I have been waiting for at least a week from them why I can not do this.
In the next topic I will tell you about interesting statistics on SSD deaths, returns, fixes and errors in work on manufacturers and models.
PS Favorite Habr, correct, please, display of a tag of 'code'.
PPS Accidentally found the official guide on kernel.org
PPPS Described on Intel Communities when, two weeks later, they bothered to answer my question about the impossibility of entering.
10 days ago I happened to leave the laptop with the battery in a critical condition overnight without charging. I don’t worry too much about battery life, but the blow came from the other side. In the morning, turning on the laptop in charge and turning it on, I was surprised to find that:
Operating system not found.
In IOS, the hard drive was determined. Having grabbed the available Ubuntu Live CD on a flash and armed with the command line, I prepared for the debug.
It should immediately be said that in cases of such failures it would be more convenient to use some Data Rescue Live CD, with already installed diagnostic utilities instead of a completely unnecessary office suite, but nonetheless.
We will recruit an arsenal, which will run for us:
$ sudo apt-get install hdpam partx smartmontools
We look what happened with us:
$ sudo partx -s / dev / sda
partx: / dev / sda: failed to read partition table
So, you can say goodbye to the partition table.
$ sudo smartctl -s on -d ata -A / dev / sda -T verypermissive
smartctl 5.41 2011-06-09 r3365 [i686-linux-3.0.0-12-generic] (local build)
Copyright © 2002-11 by Bruce Allen, smartmontools.sourceforge.net
')
SMART support is: Unavailable - device lacks SMART capability .
=== START OF ENABLE / DISABLE COMMANDS SECTION ===
Error SMART Enable failed: Input / output error
I / O Error? Disk does not support SMART? Already some nonsense.
$ sudo hdparm -I / dev / sda
/ dev / sda:
ATA device, with non-removable media
Model Number: INTEL SSDSA2CW080G3
Serial Number: BAD_CTX 00000150
Firmware Revision: 4PC10302
...
Configuration:
Logical max current
cylinders 16383 16
heads 16 16
sectors / track 63 63
-
CHS current addressable sectors: 16128
LBA user addressable sectors: 156301488
LBA48 user addressable sectors: 156301488
Logical Sector size: 512 bytes
Physical Sector size: 512 bytes
device size with M = 1024 * 1024: 76319 MBytes
device size with M = 1000 * 1000: 80026 MBytes (80 GB)
Yeah. You can see that the number of conditional SSD cylinders dropped 10,000 times and, according to the desktop Gparted, the size of the hard drive is 8MB (I confess, the console command and its output for viewing this outrage are not preserved, please take my word for it). The serial number is missing and instead BAD_CTX is something there. Well, the symptoms understood, you can refer to the search and in support. Indeed, it turns out that the problem is far from single , but, alas, I am the only one with Linux.
In short, for those who are not familiar with the language and lazy, forum users talk about the general exposure of all Intel SSDs to such a bug, especially affecting the 320 series and X25M. There is news about firmware 0362, which is designed to get rid of this particular bug, but the number of people who have already addressed this firmware with the same symptoms indicates an unsolved problem. Yes, the best solution in this case would be to send the hard drive back to Intel, so that they have an incentive to correct their mistakes.
Unfortunately, Intel support is not very efficient, and it responds about once a day, blushing on technical issues, and highly recommending installing their SSD Toolbox to determine the problem. I would like to separately note that the main slice of SSD users is the owners of MacBooks, which, similarly to me, have difficulty installing software under Windows. Special mention is worthy of the fact that this tool, designed to identify faults, requires:
- java
- .NET 3.5
- Windows Media Player Redistributable 11
which makes installing it on a computer loaded with a Live CD an almost impossible task (firstly because of the limitations of the volume on the virtual hard drive, and secondly due to the fact that WMP 11 requires verifying the authenticity of Windows, which, with creaks and groans only Some particularly prominent personalities are obtained in Wine.
Ardent greetings to the developers of this software.
I miraculously managed to explain the situation to the support, and they agreed to a replacement, but to replace, you need to fill in an incredible number of forms, to which you also need to attach confirmation of the device purchased by me. By the will of fate, I am now ten thousand kilometers from home, and did not expect such a trick.
Fortunately, on the forums, everyone unequivocally says that the contents of the disc cannot be restored, but that it is possible to restore the performance. And the time that was spent on correspondence with the support service, I did not spend in vain, but usefully spent reading forums and experiments, a brief summary of which is here and bring.
It is necessary to restore the number of cylinders, returning the cherished 16383.
For this operation, we will need two commands, the launch of both is difficult to protect against a fool and a pest.
We expose the user and password for master operations on the disk.
$ sudo hdparm –user-master user –security-set-pass abc / dev / sda
Next, we need to unlock the extended set of ATA commands, in particular secure-erase, which are blocked when the system boots. There are several ways to do this, one of which is the external box, turn it off and turn it on. I did not have an external box, but miraculously the sending of the laptop to sleep and awakening works.
The following commands perform some kind of safe cleanup, I ran both, since there was no certainty which one would be needed. Before each run the installation of the master password and closing the lid of the laptop.
$ sudo hdparm –user-master user –security-erase abc / dev / sda
$ sudo hdparm –user-master u –security-erase-enhanced abc / dev / sda
Now completely for not the faint of heart. Resetting disk settings to factory settings. To launch a command, you need another key that the command line tells you, and to clear my conscience I will not give it here, just as I will mention that this command is marked as SPECIALLY DANGEROUS and DO NOT RUN in the documentation.
$ sudo hdparm --dco-restore / dev / sda
Reboot. Hooray, you can put the system.
Total conclusions that I made for myself:
- keep your live cd at hand
- do not leave the laptop completely without power on a critical charge
- make backups, including keyrings, lists of installed packages, configs and rsa keys
- update the firmware (after you know that it works quite well)
- take care of your nerves
I want to additionally note that this method is not always fully restores functionality, and that sometimes the disk remains buggy, braking.
Once again, I send a heartfelt greeting to the support of Intel, and inform them that I still can not log in with my username and password to their community to publish this miraculous recipe from them, and remind you that I have been waiting for at least a week from them why I can not do this.
In the next topic I will tell you about interesting statistics on SSD deaths, returns, fixes and errors in work on manufacturers and models.
PS Favorite Habr, correct, please, display of a tag of 'code'.
PPS Accidentally found the official guide on kernel.org
PPPS Described on Intel Communities when, two weeks later, they bothered to answer my question about the impossibility of entering.
Source: https://habr.com/ru/post/133915/
All Articles