Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
RAID array broken, can't boot
View unanswered posts
View posts from last 24 hours

Goto page 1, 2, 3, 4  Next  
Reply to topic    Gentoo Forums Forum Index Kernel & Hardware
View previous topic :: View next topic  
Author Message
ExecutorElassus
Veteran
Veteran


Joined: 11 Mar 2004
Posts: 1185
Location: Stuttgart, Germany

PostPosted: Fri Feb 08, 2019 1:05 pm    Post subject: RAID array broken, can't boot Reply with quote

I have three drives in a raid5 array. One is dying. I pulled the plug on the wrong one and rebooted. I turned off again, plugged in the good drive, unplugged the bad. The good one was dropped, and with tge bad unplugged i couldn't boot. Turned off again, plugged the bad drive back in, booted, and it started to recover the array. After two hours, it stopped. So I tried rebooting.

Now, the initram won't boot, saying that "/dev/md127 is assemvled from 2 drives and 1 rebuilding - not enough to start the array while not clean - consider --force." It drops to a shell and says "can't access tty; job control turned off."

The keyboard won't respond.

How can I recover? I have a very old gentoo recovery CD and a USB cd drive. Are there any tools there I could use to fix thus?

The initram is from Neddy Seagoon. Neddy, can you save me once more?

Thanks in advance,

EE
Back to top
View user's profile Send private message
ExecutorElassus
Veteran
Veteran


Joined: 11 Mar 2004
Posts: 1185
Location: Stuttgart, Germany

PostPosted: Fri Feb 08, 2019 3:31 pm    Post subject: Reply with quote

update: the old SystemRescueCD I had wouldn't boot; after a few seconds of startup the machine switched off.

So now I'm burning a gentoo install CD, and hoping I can use that to recover. Or I also have a newer SystemRescueCD.

So, to be clear, I have:

sda - good
sdb - dead
sdc-old - good, but out of sync with the array
sdc-new: a new, blank HDD which I haven't plugged in yet.

How can I recover the array with the tools I have available?

Thanks for the help,

EE
Back to top
View user's profile Send private message
ExecutorElassus
Veteran
Veteran


Joined: 11 Mar 2004
Posts: 1185
Location: Stuttgart, Germany

PostPosted: Fri Feb 08, 2019 5:49 pm    Post subject: Reply with quote

UPDATE: I booted into a gentoo Live/Install CD, and now I can get to a prompt. 'cat /proc/mdstat' sows all three arrays, all inactive (and mis-numbered), and each with only two members (sdaN and sdcN).

At this point I can't start the arrays. If I type 'mdadm -A /dev/md126' I get the error that the array is not in the config file. If I give the command 'mdadm -E --scan > /etc/mdadm.conf' I then see three lines (for md1, md126, and /dev/md/carrier, the last of which is wrong).

Using 'mdadm -A /dev/md1 --force' does not work.

How do I recover from this?

Again: sdaN and sdcN are all assumed good. I have set the partition table for sdb and can add it back into the arry, but theliveCD won't do it, and I still can't boot.

thanks for the help,

EE
Back to top
View user's profile Send private message
ExecutorElassus
Veteran
Veteran


Joined: 11 Mar 2004
Posts: 1185
Location: Stuttgart, Germany

PostPosted: Fri Feb 08, 2019 6:56 pm    Post subject: Reply with quote

UPDATE2: OK, I just rewrote the array entries in /etc/mdadm.conf, and then it assembled two of them. I was able to add the appropriate sdbN partitions to those arrays, and it synced.

My last problem is the final array. This is a large array with a volume group inside with multiple partitions, holding all the rest of the system. I thought I might try a last-ditch command, and use

Code:
mdadm --create /dev/md127 --level=5 --assume-clean --metadata=1.2 --rai-devices=3 /dev/sda4 missing /dev/sdc4


but I got two errors:
that sda and sdc appear to be partof a raid array (theoldone)
and, more worrisome:
that a partition table exists on /dev/sda4 but will be lost or meaningless after creating array

How should I proceed? Can I safely ignore the warning about the partition table? How else might I be ble to rescue the array?

Thanks,

EE
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 43542
Location: 56N 3W

PostPosted: Fri Feb 08, 2019 8:35 pm    Post subject: Reply with quote

ExecutorElassus,

Do not use --create. Defaults have changed. Your data will still be there but you won't find it.
Been there, done that, don't you do it too.
If you really really want to do that (you don't yet) you must specify all the parameters to --create

What does
Code:
mdadm -E /dev/sd[abc]4
say?
Post it all, you will be glad you did later.

Post the content of /proc/mdstat too.

Read RAID Recovery and its references.

Play about with some USB drives and overlay filesystems before you trash your real data.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
ExecutorElassus
Veteran
Veteran


Joined: 11 Mar 2004
Posts: 1185
Location: Stuttgart, Germany

PostPosted: Fri Feb 08, 2019 8:46 pm    Post subject: Reply with quote

Hi Neddy,

one thing: if I try to boot with the bad drive still in, it won't assemble because it says it has two drives and one rebuilding, not enough to start, and suggests using --force. But the keyboard on the initram doesn't work. Should I just avoid that route altogether?

I can post the mdadm -E information once I boot up again from the liveCD (though I have to copy it by hand).

Stay tuned,

EE
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 43542
Location: 56N 3W

PostPosted: Fri Feb 08, 2019 8:58 pm    Post subject: Reply with quote

ExecutorElassus,

Don't copy it by hand. That post will be your get-out-of-jail free card if you run a --create and all your data vanishes.

The raid ver 1.2 metadata is written to the start of all the volumes in the raid set.
Trashing it is is free, you can do it as many times as you like without harming your data. However, once you run a --create and mess up, you have lots of parameters to get right all at the same time to find your data again.

With software raid, you can move the drives to another system and the raid set will work as well as it ever did.
It need not boot or use the raid you want to poke about on.

Don't use force yet. We need to see the mdadm -E output before you do anything you can't undo.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
ExecutorElassus
Veteran
Veteran


Joined: 11 Mar 2004
Posts: 1185
Location: Stuttgart, Germany

PostPosted: Fri Feb 08, 2019 9:11 pm    Post subject: Reply with quote

mdadm -E:
Code:

/dev/sda4;
    Magic: a92b4efc
    Version: 1.2
    feature Map: 0x0
    Array UUID: d4se5336:b75b0114:a502f2a0:178afc11
    Name: domo-kun:carrier
    Creation Time: Wed Apr 11 00:10:50 2012
    Raid Level: raid5
    raid Devices: 3

    Avail Device Size: 1931841384 (921.17 GiB 989.10 GB)
    Array Size: 1931840512 (1842.35 Gib 1978.20 GB)
    Used Dev Size: 1931840512 (921.17 GiB 989.10GB)
    Data Offset: 2048 sectors
    Super Offset: 8 sectors
    Unused Space: before=1968 sectors, after=872 sectors
    State: active
    Device UUID: 4a8d21e3:15026b07_bfacaedc:b5160599

    Update Time: fri Feb 8 12:13:10 2019
    Checksum: 8b9a3dbd - correct
    Events: 1340931

    Layout: left-symmetric
    Chunk Size: 512k

    Device Role: Active device 2
    Array State: AAA ('A' == active, '.' == missing,  'R' == replacing)

/dev/sdc4:
    Magic: a92b4efc
    Version: 1.2
    feature Map: 0x2
    Array UUID: d4se5336:b75b0114:a502f2a0:178afc11
    Name: domo-kun:carrier
    Creation Time: Wed Apr 11 00:10:50 2012
    Raid Level: raid5
    raid Devices: 3

    Avail Device Size: 1931841384 (921.17 GiB 989.10 GB)
    Array Size: 1931840512 (1842.35 Gib 1978.20 GB)
    Used Dev Size: 1931840512 (921.17 GiB 989.10GB)
    Data Offset: 2048 sectors
    Super Offset: 8 sectors
    Recovery Offset: 1389843520 sectors
    Unused Space: before=1768 sectors, after=872 sectors
    State: clean
    Device UUID: 9a99d7ad:9b5a9b75:42cb3258:cfb40e04

    Update Time: fri Feb 8 12:13:10 2019
    Checksum: fdedb47b - correct
    Events: 1340931

    Layout: left-symmetric
    Chunk Size: 512k

    Device Role: Active device 2
    Array State: AAA ('A' == active, '.' == missing,  'R' == replacing)

'cat /proc/mdstat' outputs:
Code:
Personalities: [raid6] [raid5] [raid4] [raid0] [raid1] [raid10] [linear] [multipath]
md125: active raid1 sdb3[1] sdc3[0] sda3[2]
          9765504 blocks [3/3] [UUU]

md126: active (auto.read.only) raid1 sdb1[1] sda1[2] sdc1[0]
          97536 blocks [3/3] [UUU]
So those two are fine, and it's just the problem with sdc4 being set as "rebuilding" and my not knowing how to reassemble it.

Thanks for the help,

EE
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 43542
Location: 56N 3W

PostPosted: Fri Feb 08, 2019 9:33 pm    Post subject: Reply with quote

ExecutorElassus,

We need the mdadm -E from the other volume too.

Both those volumes say
Code:
    Update Time: fri Feb 8 12:13:10 2019
    Checksum: 8b9a3dbd - correct
    Events: 1340931
so they appear to be self consistent.

That suggests that /dev/sdb4 is being rebuilt from these two.
However, only /dev/sdc4 says
Code:
Recovery Offset: 1389843520 sectors
so maybe not.
If its /dev/sdc4 that is being rebuilt, you don't have enough data to usefully do anything.

If you have a spare drive, recover /dev/sdb4 with ddrescue to the spare drive.
Depending on how that goes, it might be good enough to bring up the array.
ddrescue is surprisingly good at what it does.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
ExecutorElassus
Veteran
Veteran


Joined: 11 Mar 2004
Posts: 1185
Location: Stuttgart, Germany

PostPosted: Fri Feb 08, 2019 9:43 pm    Post subject: Reply with quote

OK, some confusion.

I had three drives in the array. sda, sdb, and sdc. sdb was dying, so I'll call id sdb-old. I have its replacement, sdb-new. Here's the timeline:

1) I unplug sdc in error, and boot
2) I see I've unplugged the wrong HDD. Shut down, plug it back in, unplug sdb-old, reboot
3) boot fails, because sdc was kicked from the array, and sdb-old is unplugged. shut down, plug in sdb-old, boot
4) recovery starts on sdc, using sda and sdb-old
5) I nod off, the screensaver kicks in, recovery stops. I shut down, try to reboot. Reboot fails because recovery is not complete
6) pull sdb-old, plug in sdb-new, boot with a liveCD
7) by renaming the arrays that mdadm wrote to /etc/mdadm.conf, I can re-start md125 and md127. I add the appropriate partitions from sdb-new, and they rebuild and are fine
8( I cannot, however, assemble md126. sda4 is fine, but sdb4 is empty so I don't want to add it yet, and sdc4 still shows as being "rebuilding". So two drives are out, and the array can't start.

So how can I proceed here? I'm not sure why the initramfs won't let me use the keyboard (it's USB) but maybe I could rebuild the array with sdb-old from the liveCD?

Cheers,

EE
Back to top
View user's profile Send private message
ExecutorElassus
Veteran
Veteran


Joined: 11 Mar 2004
Posts: 1185
Location: Stuttgart, Germany

PostPosted: Fri Feb 08, 2019 10:05 pm    Post subject: Reply with quote

UPDATE3:I tried putting sdb-back in the machine and completing recovery from the liveCD, but it failed. So I'm back where I was. Should I try again to recover? It slowed way down at about 80%, then sdb(old)4 got set F, sdc4 got set S, and the recovery stopped and the array went inactive.

What now?
EE


Last edited by ExecutorElassus on Fri Feb 08, 2019 10:32 pm; edited 1 time in total
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 43542
Location: 56N 3W

PostPosted: Fri Feb 08, 2019 10:16 pm    Post subject: Reply with quote

ExecutorElassus,

Quote:
1) I unplug sdc in error, and boot

This brings up the array in degraded mode on sd[ab]4

Quote:
2) I see I've unplugged the wrong HDD. Shut down, plug it back in, unplug sdb-old, reboot
3) boot fails, because sdc was kicked from the array, and sdb-old is unplugged. shut down, plug in sdb-old, boot

This won't start as sd[ac]4 now have different event counts. You said that.

At the outset sd[abc]4 were self consistent.

Other than the rebuild (which should do nothing on a self consistent set) what writes were there when sdc4 was unplugged?
If you are 100% sure there were none, then force is safe. If you get that wrong, horrible things will happen.

Heres the safe approach.
Partition the replacement sdb. Do not try to do anything with md126
Use ddrescue to image the dying sdb4 onto its replacement. Post the ddrescue log.
If ddrecue is successful, sda4 and the new sdb4 can continue to rebuild sdc4.
This will be a bit of a mess as you will have three raid sets spread over four drives.

While ddrescue does its thing, read up on the mdadm --replace command.
What you should have been doing is add the new drive to the system.
Partition the new drive, then with all the raids working on three drives, --replace the sdb elements.
When replace completes, it drops the sdb element out of the array.
Replace has the advantage of building the new element from any n-1 of the n elements in the array, which can vary as the replace progresses.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 43542
Location: 56N 3W

PostPosted: Fri Feb 08, 2019 10:17 pm    Post subject: Reply with quote

ExecutorElassus,

No, don't pull the bad one until its been replaced.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
ExecutorElassus
Veteran
Veteran


Joined: 11 Mar 2004
Posts: 1185
Location: Stuttgart, Germany

PostPosted: Fri Feb 08, 2019 10:37 pm    Post subject: Reply with quote

Hi Neddy,

I'm certain there was nothing going on. I booted, started the array recovery, started the wm, and then recovery stopped (see above).

How do I find ddrescue? It's not on the liveCD. And I'm going to have to find an extra set of cables, as I think I'm full.

But I'm done for tonight. Have to sleep.

Cheers,

EE
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 43542
Location: 56N 3W

PostPosted: Fri Feb 08, 2019 10:45 pm    Post subject: Reply with quote

ExecutorElassus,

dmesg will tell why the recovery stopped.
Maybe a read error?

The WM will write files.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
ExecutorElassus
Veteran
Veteran


Joined: 11 Mar 2004
Posts: 1185
Location: Stuttgart, Germany

PostPosted: Sat Feb 09, 2019 8:16 am    Post subject: Reply with quote

I tried to re-start recovery by booting a liveCD and re-starting the array with sdb-old. It resumed from 74%, but then at about 80% the throughput started slowing down considerably, then sdb-old was set as (F) and sdc set (S) and both dropped from the array and the array set inactive.

So what now? try to re-start recovery? Try to use ddrescue with both sdb-old and sdb-new plugged in?

Thanks for the help,

EE
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 43542
Location: 56N 3W

PostPosted: Sat Feb 09, 2019 11:29 am    Post subject: Reply with quote

ExecutorElassus,

The 80%, where sdb-old was marked as (F) is where ordinary retries failed and the kernel gave up reading the drive.

You need to try harder to read sdb-old. That's what ddrescue does. Only sdb4-old though.
Use ddrescue to image sdb4-old onto sdb4-new.

If you take both sdb-old and sdb-new to another machine, you can image all of sdb and the other raid sets will adopt it as their own when you put sdb-new back in place of sdb-old.

Once sdb4-old is imaged onto sdb4-new the raid rebuild will continue.
We will need to check the ddrescue log becaue and data not recovered will be rubbish but the kernel will happily read that and build it into your raid rebuild.
Its essential that you squeeze ddrescue until the pips squeek.

ddrescue tries much harder than the kernel to get your data off the drive, so it way well work where the kernel fails. It will get most of the data as fast as the kernel, then it will slow down.
With only a few blocks to go, I've seen in as slow as 1 physical block per hour..
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
ExecutorElassus
Veteran
Veteran


Joined: 11 Mar 2004
Posts: 1185
Location: Stuttgart, Germany

PostPosted: Sat Feb 09, 2019 1:04 pm    Post subject: Reply with quote

Hi Neddy,

ddrescue has been running now for two hours, reporting 1h15m left to go. "Current rate" usually stays around 110MB/s, but then stops for a few seconds every 10s or so. The "average rate" has been steadily dropping from 120MB/s down to 70MB/s now. I'm at 53% complete.

the command I used was:
Code:
ddrescue -f -n /dev/sdb4 /dev/sdd4 /root/rescue.log


Once that completes, the guide tells me I should use
Code:
ddrescue -d -f -r3 /dev/sdb4 /dev/sdd4 /root/rescue.log

to attempt to recover any failures.

Assuming this completes (hopefully just the first, but when not, both), what is the next step?

If it keeps slowing down, can/should I really just let it keep running?

Thank for the help,

EE
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 43542
Location: 56N 3W

PostPosted: Sat Feb 09, 2019 2:00 pm    Post subject: Reply with quote

ExecutorElassus,

The slowdown is a feature of HDD, not ddrescue.

Rotating rust HDD are 'zoned'. The number of sectors per physical track is not a constant, so the read rate is higher near the edge of the platter that near the spindle.
A factor of 2x or more is common. Each zone has fewer sectors per track, so there is less data per revolution of the platter.

The file /root/rescue.log is a text file. You can read it any time.
A + at the end of the line is recovered.
A - at the end of the line indicates a problem.

From memory, (I don't have a sample to hand), the columns are start, end and size in hex bytes.
Size is always an integer multiple of the physical block size. Thats 0x200 for 512B blocks, or 0x1000 for 4kB blocks.

ddrescue is just dd with some extra tricks. The extra tricks are only used as required.
When ddrescue stops, post the log. We will tell it to try harder and use a few more tricks that ddrescue can't do unaided.

When you rerun ddrescue with an existing log file, it knows what has already been recovered and won't try that data again.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
ExecutorElassus
Veteran
Veteran


Joined: 11 Mar 2004
Posts: 1185
Location: Stuttgart, Germany

PostPosted: Sat Feb 09, 2019 3:55 pm    Post subject: Reply with quote

Hi Neddy,

ddrescue finished. Here's what its status says:
Code:
GNU ddrescue 1.21
Press Ctrl-C to interrupt
       ipos:   971398,  MB non-trimmed:       0 B,     current rate: 5461   B/s
      opos:   971398  MB  non-scraped: 1000  kB,   average rate:   57412   kB/s
non-tried:             0   B   errsize:          36352  B      run time: 4h 47m   8s
rescued:    989102   MB,  errors:               71, remaining time:              3m
percet rescued:    99.99%             time since last successful read:      0s
Finished
zsh: bus error   ddrescue -f -n /dev/sdb4 /dev/sdd4 /root/rescue.log

/root/rescue.log reads like this:
Code:
# Mapfile. Created by GNU ddrescue version 1.212
# Command line: ddrescue -f -n /dev/sdb4 /dev/sdd4 /root/rescue.log
# Start time: 2019-02-09  10:56:52
# Current time: 2019-02-09 15:44:00
# Finished
# current_pos   current_status
0xE22BDDC000     +
#       pos               size     status
0x00000000   0xAAA4A18000   +
0xAAA4A18000 0x00000200     -
0xAAA4A18200 0x00000C00    /
0xAAA4A18E00 0x00000200    -
0xAAA4A19000 0x000BC000  +


That goes on for wuite a while. Do you want me to copy that all out?

what should I do now?

Thanks,

EE
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 43542
Location: 56N 3W

PostPosted: Sat Feb 09, 2019 4:18 pm    Post subject: Reply with quote

ExecutorElassus,

Code:
#       pos               size     status
0x00000000   0xAAA4A18000   +
0xAAA4A18000 0x00000200     -
0xAAA4A18200 0x00000C00    /
...


The first line tells of recovered data from the beginning of the drive. That's good.
The second line tells of a single disk block not yet recovered.
The third line tells of 6 consecutive blocks that ddrescue didn't try very hard with. 0xC00/0x200.

Make sure you can handle the failing drive while ddrescue runs. The idea now is to tell ddrescue to try harder while you operate the drive the right way up, upside down and on each edge in turn.
Its OK to move the drive slowly while it operates.

Lets look at the options ...
Code:
       -d, --idirect
              use direct disc access for input file
thats good.
Code:
      -r, --retry-passes=<n>
You used retry=3 ... 256 is better, especially as you will move the drive between six orientations.

Code:
       -A, --try-again
              mark non-trimmed, non-scraped as non-tried
That's like a reset between all 256 attempts. Add that.

You have
Code:
errsize:          36352  B
more bytes to get. That's decimal, not hex.

If its still not got all the data after the above, add in
Code:
       -R, --reverse
              reverse the direction of all passes
so it starts from the end of the drive. This is really slow, as it adds a lot of latency.


Code:
# Command line: ddrescue -f -n /dev/sdb4 /dev/sdd4 /root/rescue.log


You told it not to try too hard with -n.
Code:
ddrescue -f -r 256 -d -A /dev/sdb4 /dev/sdd4 /root/rescue.log
looks good and try all 4 edges and both faces while that runs.
It its still not got all your data, add -R to the above command and do it again.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
ExecutorElassus
Veteran
Veteran


Joined: 11 Mar 2004
Posts: 1185
Location: Stuttgart, Germany

PostPosted: Sat Feb 09, 2019 4:41 pm    Post subject: Reply with quote

All right, now I'm running the command
Code:
ddrescue -d -f -r20 -A /dev/sdb4 /dev/sdd4 /root/rescue.log

it's presently scraping failed blocks. The errsize is 163 kB, errors is 118, and it says it has 10 hours to go. Last successful read goes up to about 30s, sometime 3m, before going back to 0.

Obviously, I won't be able to stand here and rock the HDD back and forth for 10 hours (and since the cable-salad in the machine is kinda dens, it's not easy anyway).

Am I doing OK so far here?

Cheers,

EE
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 43542
Location: 56N 3W

PostPosted: Sat Feb 09, 2019 4:54 pm    Post subject: Reply with quote

ExecutorElassus,

Code:
non-tried:             0   B   errsize:          36352  B      run time: 4h 47m   8s

That's only 36kB of error. It should never increase.
Are you copy typing inaccurately?

-r20 is OK, turn the drive every time ddrescue stops and rerun the command.
Its only trying hard/impossible to read sectors now, so its going to be slow.
There is no need to rock the drive continuously. I use a large -r and turn the drive from time to time.

Every pass in the -r20 tries all the unread areas only. All that matters is that the errsize reduces.
The error count can increase, if it gets a block out the middle of a larger block of errors.

When you go away for the night, switch the system off and let the failed drive cool.
In the morning, you may find the cool drive reads better.
Do not put the drive in the fridge/freezer to aid cooling.

All you are trying to do is to coax one last read of the 'failed' sectors.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
ExecutorElassus
Veteran
Veteran


Joined: 11 Mar 2004
Posts: 1185
Location: Stuttgart, Germany

PostPosted: Sat Feb 09, 2019 4:59 pm    Post subject: Reply with quote

Hi Neddy,

the center column now reads:
Code:
non-trimmed: 0       non-scraped:     357376 [this number is decreasing]     errsize:   366  kB  [this number is increasing]   errors:    195


is that bad?

To shut down for the night, can I just do 'init 0' from another TTY? Will it pick up again in the morning where it left off if the Mapfile disappears? (this is a liveCD)

Cheers,

EE
Back to top
View user's profile Send private message
NeddySeagoon
Administrator
Administrator


Joined: 05 Jul 2003
Posts: 43542
Location: 56N 3W

PostPosted: Sat Feb 09, 2019 5:07 pm    Post subject: Reply with quote

ExecutorElassus,

The log file must persist from run to run, or ddrescue will start over again.
That's not what you want/need with a dying drive.

You can save the log, presently, /root/rescue.log to a USB stick and keep it there for future runs.
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Kernel & Hardware All times are GMT
Goto page 1, 2, 3, 4  Next
Page 1 of 4

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum