Server Version#: 1.32.5.7349
Player Version#: Version 1.74.1.3913-9baac009
<If providing server logs please do NOT turn on verbose logging, only debug logging should be enabled>
I’m having a really infuriating issue with my PMS. I just moved over to unRAID with Docker and rebuilt my entire server from scratch multiple times. I have tried Binhex Plex, the official PMS docker, and am now using LinuxServers. This issue has now happened on every single one of them
It will work fine for a couple of days and then it will always revert to the same behavior. First, any newly added media will not be detected, and then I completely lose the ability to edit any of my libraries. I cannot modify the folders for the library, the agent, or any of the metadata for any of my library items. I can’t even delete the specific libraries to read them. To me it’s really odd how it will work flawlessly and then randomly decide to cut out
I have read all of the other threads about this and none of the suggestions work. I cannot keep rebuilding my entire library and Docker container every few days to remedy this. I need to reliably know that my media will show up when it is added and that I can modify the libraries.
If anyone has any insight I would really appreciate it!
Here are the logs:
Plex Media Server Logs_2023-07-31_21-00-19.zip (1.3 MB)
Your database is corrupt. I don’t know what would keep causing that over and over on fresh installs
ERROR - [DatabaseFixups] SQLITE3:0x80000001, 11, database corruption at line 84326 of [a29f994989]
Cannot commit transaction. database disk image is malformed
Bad Cache drive or bad ram seem to be the likely suspects from a quick Google search
do you have any ideas on why this would keep happening? This is on a brand new SSD and brand new RAM. This is the second unRAID server I have had this on as well.
I set this screen update the database size from the default 40MB to 128MB. Could this have done it?
I just restored the database to a previous version and now it works fine, but it’s odd this has happened multiple times on multiple separate computers with multiple different Docker containers. any ideas what could be causing this?
My Plex app data lives on Cache only. I have 2 980 Pro’s in BTRFS in RAID1
Turn down the database cache size.
If you turn it up to high, you can get lots of corruption.
It’s best to use the default value
What does this mean (implementation wise)?
My Plex app data lives on Cache only.
@LoyalSnoop
If I may suggest?
-
Samsung 980 Pro SSDs have a bad firmware bug. Make sure you’ve updated them to latest else you’ll be plagued with problems.
-
Don’t use BTRFS. It won’t cut it here. Use XFS formatting (it handles the throughput requirement far better)
-
RAID 1 (mirror) only burns out both SSDs at the same time – Why waste it?
Minimally, the 980 Pro SSD SATA is 535 MB/sec, and NVMe is minimum 2+ GB/sec on PCIe 3.0.
- observe PCIe NVMe software RAID 0 (concat) … which is slow as heck on an old xeon.
[chuck@glockner ~.2001]$ df
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/md0 243421052 33111452 210309600 14% /
/dev/md1 976032144 5712712 970319432 1% /home
/dev/md2 1999168256 949021056 1050147200 48% /vmssd
/dev/sda1 523248 6216 517032 2% /boot/efi
/dev/md3 117185445888 58217479204 58967966684 50% /vol
[chuck@glockner ~.2002]$ sudo dd if=/dev/md2 of=/dev/null bs=4M count=10000 status=progress
41720741888 bytes (42 GB, 39 GiB) copied, 20 s, 2.1 GB/s
10000+0 records in
10000+0 records out
41943040000 bytes (42 GB, 39 GiB) copied, 20.1124 s, 2.1 GB/s
[chuck@glockner ~.2003]$ sudo blkid /dev/md2
/dev/md2: LABEL="vmssd" UUID="d9603f57-6824-4a6e-81e4-edf64f654634" BLOCK_SIZE="512" TYPE="xfs"
[chuck@glockner ~.2004]$
My server is HDD based and suffers no performance (search or other DB) issues.
Have you used my DB optimize tool?
thanks for your feedback.
Database size has been set back to default of 40MV
I’m not worried about the cost of the SSD’s. The redundancy is what I was after.
I’ll look into updating my Samsung firmware. I was unaware of the bug, as I thought it only affected the 990 series
Is XFS sufficient for a cache SSD? I had heard that it doesn’t handle TRIM very well.
Could you explain a little bit more about the pros and cons of XFS and BTEFS?
And also no I haven’t tried your tool yet. Where can I find it?
PS I hope you get better soon’
All the 980 SSDs are flawed. No getting around it.
XFS is ideal.
Enterprise servers (web servers) are big SSD + XFS -or- SSD + ZFS.
I use it on everything here. My NUC8-i7-HVK (the gaming NUC) is my Linux workstation/desktop- all formatted XFS.
I avoid BTRFS because of all the horrors I helped folks through on Synology.
I can’t tell you how many entire HDD volumes were lost to BTRFS (which itself is a RAID). I do not understand, because there is no purpose, to put a RAID filesystem on top of a RAIDed storage volume (double RAID layers makes it slow and wasteful).
The tool is here.
Check the README to see how to use it best.
Typically -
- Put inside the container where /config is
- Get into the container (docker exec -it container bash)
- chmod +x DBRepair.sh
- ./DBRepair.sh
– Stop. Auto. Start. Exit
– (should just about do everything you need)
Be advised, Binhex doesn’t support start/stop. I can stop Plex but couldn’t make start work correctly like I can everything else.
Thanks for the info. next time this issue arises i’ll ask you for help and try your tool
Please do update the firmware on your SSDs before they lock up.
The bug with 980 SSDs is they go READ-ONLY around the 2 TBW mark.
do you know if upgrading them wipes them?
do you know a way do upgrade the samsung firmware in UNRAID?
@LoyalSnoop
The firmware update will not erase the SSD.
I upgraded the 980’s in my pfsense without issue.
Thanks for all your insight. I’ll be upgrading the SSD firmware today when I get home from work
I’ll try to update natively in docker, if not I’ll boot to Hiren’s boot CD and update from there
I think you missed a point.
- Download the firmware update ISO image from Samsung for your SSD
- Burn that onto a USB thumb drive
- Restart device and boot from thumb drive
- Samsung firmware update console (text) will show
- It will scan and then confirm what to update.
– It won’t update devices which don’t match.
Please do read their instructions
just got home from work and went to update ssd firmware
i am already on the latest version
do you have any other suggestions?
also my BIOS says I have a GPT header corruption on one of my drives, but the logs don’t show it in unRAID and the SMART test comes back fine. do you have any input on this?
ignore GPT header error. it’s caused by my cheap SATA card. i have a SAS card arriving in the mail on thursday
i would appreciate your input on the other stuff though
I also use a SAS/SATA card for my system with SATA drives without issue. I use the 9201 LSI. Looks like this with lshw.
[chuck@glockner ~.2010]$ sudo lshw -C storage
*-sas
description: Serial Attached SCSI controller
product: SAS2116 PCI-Express Fusion-MPT SAS-2 [Meteor]
vendor: Broadcom / LSI
physical id: 0
bus info: pci@0000:01:00.0
logical name: scsi0
version: 02
width: 64 bits
clock: 33MHz
capabilities: sas pm pciexpress vpd msi msix bus_master cap_list rom
configuration: driver=mpt3sas latency=0
resources: irq:24 ioport:e000(size=256) memory:93280000-93283fff memory:93080000-930bffff memory:93000000-9307ffff memory:93284000-9329ffff memory:930c0000-9327ffff
Regarding the SSDs,
Good that you’re up to date on firmware.
You have them in RAID 1?
Still formatted BTRFS ?
awesome. i got a used LSI card as well!
was it plug and play for you or did you need to do any config? mine is coming pre flashed in IT mode
