PMS Pagefaults - UNRAID - Docker

[urbanracer34]

I also have the same problem. I get segfaults in my system log and it is only from Plex. At least once a day.

Feb 16 02:05:40 GIBSON kernel: Plex Media Scan[6126]: segfault at 154ea3b8ffd0 ip 0000154ea8bb3152 sp 00007ffccad56e90 error 4 in Plex Media Scanner[154ea8b5f000+2dd000]
Feb 16 02:05:40 GIBSON kernel: Code: a5 70 ff ff ff 0f 84 fc 00 00 00 48 89 8d 78 ff ff ff 48 8b 01 48 63 08 48 6b c9 68 4c 8b ad e0 fd ff ff 4c 8b bd e8 fd ff ff <4a> 8b 4c 29 c0 48 89 4d 98 4c 63 70 04 49 6b c6 68 42 8b 5c 28 c0
Feb 16 03:00:16 GIBSON crond[1203]: exit status 1 from user root /usr/local/sbin/mover &> /dev/null

[urbanracer34]

Still ongoing!

Feb 16 02:05:40 GIBSON kernel: Plex Media Scan[6126]: segfault at 154ea3b8ffd0 ip 0000154ea8bb3152 sp 00007ffccad56e90 error 4 in Plex Media Scanner[154ea8b5f000+2dd000]
Feb 16 02:05:40 GIBSON kernel: Code: a5 70 ff ff ff 0f 84 fc 00 00 00 48 89 8d 78 ff ff ff 48 8b 01 48 63 08 48 6b c9 68 4c 8b ad e0 fd ff ff 4c 8b bd e8 fd ff ff <4a> 8b 4c 29 c0 48 89 4d 98 4c 63 70 04 49 6b c6 68 42 8b 5c 28 c0
Feb 16 03:00:16 GIBSON crond[1203]: exit status 1 from user root /usr/local/sbin/mover &> /dev/null
Feb 17 02:04:55 GIBSON kernel: Plex Media Scan[7246]: segfault at 148bd7065fd0 ip 0000148bdbfb3152 sp 00007ffd50580f30 error 4 in Plex Media Scanner[148bdbf5f000+2dd000]
Feb 17 02:04:55 GIBSON kernel: Code: a5 70 ff ff ff 0f 84 fc 00 00 00 48 89 8d 78 ff ff ff 48 8b 01 48 63 08 48 6b c9 68 4c 8b ad e0 fd ff ff 4c 8b bd e8 fd ff ff <4a> 8b 4c 29 c0 48 89 4d 98 4c 63 70 04 49 6b c6 68 42 8b 5c 28 c0

[ChuckPa]

Up and running. (until I restart?)

What timestamp do the core dumps report?

[urbanracer34]

Anywhere between 2:00 to 4:00 (24hr time, GMT-6) according to my unRAID system log.

The process that this happens with is Plex Media Scanner. At least in my case.

[ChuckPa]

Thanks.

I’ll presume that coincides with Scheduled Tasks (2-5 am default)

and “Deep Analysis” unless you tell me otherwise.

[urbanracer34]

This is correct.

[ChuckPa]

I ran Unraid+PMS docker last night.

The UNRAID host crashed.
It complained the thumb drive (license/OS) was invalid.
Took the whole thing down.

I spent the last hour unFing it.

I knew something was going oddly when, twice during the evening, UNRAID told me the thumb was corrupt.

All I had to do was reboot it. (I’d checked the filesystem and found it ok)

I’ll be candid – Unraid is NOT looking very good right now. I know that many swear by it. I’m still ‘swearing at it’ stage

Here’s a really stupid question. Where’s the graphical File Explorer ?
Every NAS I know of has one built in.

[urbanracer34]

You can usually see what files are on a disk in unRAID by clicking “view” on the right side of the screen beside a drive entry.

If you want something more robust, install the “Community Applications” plugin. Go to the new “apps” section on the top of the screen and install the binhex-krusader container.

That is what I use.

[urbanracer34]

@ChuckPa

One thing I don’t have indexed is music or photos. Do you have either ?

I have none of these on my server.

I honestly hope we can get to the bottom of this soon.

[ChuckPa]

Fair enough… Fingers crossed

[urbanracer34]

Just one more thought. I found a plugin called Dynamix File Explorer and it allows to directly manipulate the filesystem without needing Binhex-krusader.

[ChuckPa]

Thanks. I have a bunch of media loaded up

Now we see what happens.

If the OS blows up and requires reboot a 5th time and/or making another thumb drive – What else can be done? I won’t be able to trust the platform to figure out if PMS is the root problem.

[urbanracer34]

Sadly I don’t have the answer if unRAID keeps buying the farm.

Hopefully it lives long enough to get us an answer.

[ChuckPa]

As a fellow NAS owner, I’m also now concerned about you.

You have your data on it. If it keeps blowing up, how safe is your data?

Linux is Linux and everyone seems to boot from a thumb drive with unraid.
(which I take exception to about how to protect licensing)

[urbanracer34]

A lot of my data can be reacquired, it just would take a lot of time.

Other than that, my server is the primary repository for my family.

[ChuckPa]

I can re-rip mine too but that’s a LOT of hours I care not to waste ( i have a big stack of blurays ). I estimate it would take me about 6 months to recover from complete media loss.

The method I’ve chosen allows me to pull all the drives out and put them, in any configuration, into a Linux machine (even workstation) and reconstitute the volume. I use mdadm RAID 6)

I’ve found XFS far more robust than EXT4 after abrupt power failures.

ZFS, while an improvement, doesn’t perform as well as XFS on HDDs.
I would need 48 4TB SSDs to equal what I have with HDDs. I can’t afford that.

[ChuckPa]

@urbanracer34

It just randomly did it again.

It seems it won’t run more than 8 hours without going Toes Up

I did set it for Dark theme so I can see this happen easily.

If I reboot, I’ll find the thumb drive all messed up and needing to be fixed before it can be booted again.

[urbanracer34]

Try to swap out the USB key entirely. Totally different brand if you can. The one you are using might not be of good quality.

I have a good size UPS directly hooked to the server and to shut it down when a certain threshold is reached.

[ChuckPa]

Sandisk isn’t good enough quality ? ODD

[urbanracer34]

PSA on SanDisk USBs