Filesystem check exited with signal 9

[Дмитрий Варенов]

Good day everyone!

Guys, i need your advice. After a power outage i turned the device back on.

After booting, my /dev/md0 device was mounted in read-only mode and the device complains about inability to automaticaly fix a filesystem on /dev/md0 due to "fsck exited with signal 9"

1. I've unmounted /dev/md0 and tried to run "fsck /dev/md0" manualy using SSH connection and also tried to do the same from web-interface (Disk --> Filesystems --> FS Operations) - the result is the same - "exited with signal 9"

2 Then i've manually remounted /dev/md0 to RW mode, system becomes opertional and works as usual. After that i've tried to do "Reboot&Check" - same result, the fsck "exits with signal 9" and /dev/md0 mounted RO after reboot  

I've been using this storage for almost 10 years i guess, and last 4 or 5 of them using ALT-F. During this years the system suffered from numerous power outages, a couple of times with the FS damage but this time the problem is very strage.

Please help to find out the culprit of the problem and fix the FS checking.

Thanks in advance for all ideas and suggestions 

The system is Alt-f 1.0 running on DNS-323-A1

RAID-1 (Mirror) of two ST4000VM000-1F3168 4.0TB HDDs - EXT-4 FS 

[Bert Broekhuizen]

First thing to check is if you have swap partitions and if they are enabled.

From ssh try the following:

$ free

and

$ cat /proc/swaps

The first command will show memory usage info. The last line should look something like this (actual numbers from my DNS-323-B1):

Swap:       976552            0       976552

The first and the last number should be non-zero. If the second one is non-zero then swapping has actually taken place.

The second command will show which partitions are used as swap. Again from my DNS-323-B1:

Filename                Type        Size    Used    Priority
/dev/sda1               partition   488276  0       1
/dev/sdb1               partition   488276  0       1

If you don't have any swap partitions, then there will not be enough virtual memory available for fcsk to do the filesystem check.

Best regards,

Bert

Op zaterdag 28 juli 2018 16:20:28 UTC+2 schreef Дмитрий Варенов:

[Bert Broekhuizen]

Actually, you can also see on the status page of the web interface if swap is available and used.

Op zaterdag 28 juli 2018 17:04:28 UTC+2 schreef Bert Broekhuizen:

[Bert Broekhuizen]

Screenshot of my status page:

Op zaterdag 28 juli 2018 17:07:03 UTC+2 schreef Bert Broekhuizen:

[João Cardoso]

On Saturday, 28 July 2018 15:20:28 UTC+1, Дмитрий Варенов wrote:

Good day everyone!

Guys, i need your advice. After a power outage i turned the device back on.

After booting, my /dev/md0 device was mounted in read-only mode and the device complains about inability to automaticaly fix a filesystem on /dev/md0 due to "fsck exited with signal 9"

That is indeed very strange. Where does it appears? In the Errors/Warnings Status webUI? No more messages? You might want to watch the "System Log", System->Utilities, Logs, and search for more fsck messages and disk errors messages. Or better, post the full log.

"Signal 9" means that the fsck process was deliberately force killed.

Bert's tips are important, but lack of swap space generally generate a memory related error. And swap must be setup on the first partition disk, sda1 and sdb1.

You say that you have the system by several years now -- are the disks that old also? You might want to see the SMART status, Disk->Utilities, Health, Status (post the full output), and perform a short or long test.

Being able to manually mounting the filesystem RW doesn't mean that the filesystem is OK, fs checking does a more deep verification, and the SMART test verifies even not filesystem used disk areas.

[Дмитрий Варенов]

Thank you very much, guys!

Bert was correct. I have a swap partition but, i guess i accidentally set the swappiness value to ZERO, which used to be set to 40.

As far as i undestand zero value means "Swap disable". So i set it back to 40 and retried running "Reboot&Check".

Check seems to be running fine, now waiting for it to finish.

I've attached a screenshot showing that indeed it uses swap actively during system check.

Joao, concering the HDDs and SMART i also have some questions but i guess it's a topic for another thread :-) 

--
You received this message because you are subscribed to the Google Groups "Alt-F" group.
To unsubscribe from this group and stop receiving emails from it, send an email to alt-f+unsubscribe@googlegroups.com.
Visit this group at https://groups.google.com/group/alt-f.
For more options, visit https://groups.google.com/d/optout.

[João Cardoso]

On Saturday, 28 July 2018 19:15:14 UTC+1, Дмитрий Варенов wrote:

Thank you very much, guys!

Bert was correct. I have a swap partition but, i guess i accidentally set the swappiness value to ZERO, which used to be set to 40.

As far as i undestand zero value means "Swap disable".

 

Yes, the tooltip says:  "Swappiness specifies how aggressively swap is used, with 0 almost disabling it and 100 actively using it."

So i set it back to 40 and retried running "Reboot&Check".

Check seems to be running fine, now waiting for it to finish.

I've attached a screenshot showing that indeed it uses swap actively during system check.

Joao, concering the HDDs and SMART i also have some questions but i guess it's a topic for another thread :-) 

Yes, that's better. Be warned that the blue color means "warning", so although your disks health are OK as reported by SMART, in the past there was some issues -- probably a transient high temperature because of some fan issues or insufficient air flow due to not enough space clearance around the box.