HC1 SATA drive errors -- how to recover?

Moderators: mdrjr, odroid

HC1 SATA drive errors -- how to recover?

Unread postby ASword » Mon Aug 06, 2018 1:55 am

I've got an HC-1 with a 250GB ADATA SSD that I've been running postgresql and grafana on. A few days ago the database abruptly went down and the filesystem on the SSD became readonly. I tried rebooting but the system wouldn't come up with the SSD mounted. I've been trying various options on fsck and e2fsck, however I still can't mount the drive. It reports lots of I/O errors. Here was the latest run of e2fsck:

Code: Select all
asword@hc1:~$ sudo e2fsck -c -C 0 -p -b 32768 -k -v /dev/sda2
/dev/sda2: Superblock needs_recovery flag is clear, but journal has data.
/dev/sda2: Recovery flag not set in backup superblock, so running journal anyway.
/dev/sda2: recovering journal
/dev/sda2: Superblock needs_recovery flag is clear, but journal has data.
/dev/sda2: Recovery flag not set in backup superblock, so running journal anyway.
/dev/sda2: Superblock needs_recovery flag is clear, but journal has data.
/dev/sda2: Recovery flag not set in backup superblock, so running journal anyway.
e2fsck: unable to set superblock flags on /dev/sda2

/dev/sda2: ***** FILE SYSTEM WAS MODIFIED *****

/dev/sda2: ********** WARNING: Filesystem still has errors **********


The subsequent mount command is still sitting there trying... at least half an hour later.

Ideally I could mount the thing and recover whatever files are possible to copy off, then reformat/repartition the drive. I'm not sure what caused the sudden burst of error problems, but it could be a combination of the heat lately as well as running compiles that pegged all 8 cores for a while. I'm using the ODROID 4A 5v power supply, so its possible that I ran into some power problems.

Any tips on how to go about recovering or mounting the drive? I'm running Ubuntu 16.04 with kernel 4.14.5-92.
ASword
 
Posts: 76
Joined: Fri Aug 04, 2017 12:48 pm
languages_spoken: english
ODROIDs: XU4, HC1

Re: HC1 SATA drive errors -- how to recover?

Unread postby ASword » Tue Aug 07, 2018 4:17 am

In case anyone is interested, even though my partitions would not mount, I could use the program myrescue (a variant of dd_rescue which is a variant of dd) to copy the partition off into an image file on another volume. Particularly strange is that myrescue reported zero errors while reading the original drive. From there I could mount the file (also no problem doing so), and am now copying the contents out to be more readily accessible. Next I will wipe the original drive and try rebuilding it from scratch.
ASword
 
Posts: 76
Joined: Fri Aug 04, 2017 12:48 pm
languages_spoken: english
ODROIDs: XU4, HC1

Re: HC1 SATA drive errors -- how to recover?

Unread postby ASword » Wed Aug 08, 2018 9:46 pm

Just to round out my saga: tried everything I could find about wiping and recreating the partition table, only to come to the conclusion that something critical has failed in the SSD itself. Have replaced the ADATA unit with an Intel unit (and the 4A power supply with a 5A power supply), and hope that it is more robust.
ASword
 
Posts: 76
Joined: Fri Aug 04, 2017 12:48 pm
languages_spoken: english
ODROIDs: XU4, HC1

Re: HC1 SATA drive errors -- how to recover?

Unread postby rooted » Wed Aug 08, 2018 10:53 pm

Best of luck with the new drive.
User avatar
rooted
 
Posts: 5317
Joined: Fri Dec 19, 2014 9:12 am
Location: Gulf of Mexico, US
languages_spoken: english
ODROIDs: C1, C1+, C2
XU3 Lite, XU4
N1
VU7+
HiFi Shield 2
Smart Power (original)

Re: HC1 SATA drive errors -- how to recover?

Unread postby ASword » Thu Aug 09, 2018 2:51 am

Yeah, I'm hoping Intel's reputation is well deserved and the negative ADATA comments I've read in various places do indicate that premature failures do happen. Just goes to show you can't rely on the vendor provided MBTF numbers, which was one of my motivators in choosing that unit in the first place (2M hrs claimed, about 8K realized in my case -- supremely bad luck or systemic problem?!). I could probably go back to ADATA with a warrantly claim, but frankly at about $100 its not worth my time!
ASword
 
Posts: 76
Joined: Fri Aug 04, 2017 12:48 pm
languages_spoken: english
ODROIDs: XU4, HC1

Re: HC1 SATA drive errors -- how to recover?

Unread postby rooted » Thu Aug 09, 2018 6:11 am

Any reason you chose Intel over Samsung?
User avatar
rooted
 
Posts: 5317
Joined: Fri Dec 19, 2014 9:12 am
Location: Gulf of Mexico, US
languages_spoken: english
ODROIDs: C1, C1+, C2
XU3 Lite, XU4
N1
VU7+
HiFi Shield 2
Smart Power (original)

Re: HC1 SATA drive errors -- how to recover?

Unread postby paramedic233 » Tue Aug 14, 2018 7:02 am

You have Grafana running?
I'm trying to find Grafana for my XU4, but armhf is not supported.
Are you running an older version?
paramedic233
 
Posts: 29
Joined: Tue May 01, 2018 4:47 am
languages_spoken: english
ODROIDs: XU4Q, Cloudshell2

Re: HC1 SATA drive errors -- how to recover?

Unread postby ASword » Tue Aug 14, 2018 10:32 am

Chose Intel because they had a sale on and didn’t on the Samsung. 25% per GB difference and both have good quality controllers (I’ve heard).

Grafana: I was running my own build up until last week when I was doing a round of updates. Noticed that they do now have ARM builds, including armhf. Go check their website.
ASword
 
Posts: 76
Joined: Fri Aug 04, 2017 12:48 pm
languages_spoken: english
ODROIDs: XU4, HC1


Return to Ubuntu

Who is online

Users browsing this forum: No registered users and 4 guests