Jaster
-
Posts
409 -
Joined
-
Last visited
Content Type
Profiles
Forums
Downloads
Store
Gallery
Bug Reports
Documentation
Landing
Posts posted by Jaster
-
-
I'm running four production servers and two labs. Ramping up to six production within the next two weeks.
Data loss has never occurred - I had a lot of "bad luck" with BTRFS, but I moved away to XFS and am fine now.
unraid is VERY stable, however some "rogue" VM's and docker were able to cause some issues (by flooding the log, etc) - but that would also kill any other OS/distro with a lv1 hypervisor.
Two things are missing from my point of view:
1. some kind of reliable notifications - this is especially annoying if you fail over on WAN and don't receive notifications.
2. VM snapshots (there are some solutions, but those ain't great).
Again, that stuff does not come out of the box on other OS either.
-
Hi, Suddenly this appeared;
How do I get rid of that?
-
40 minutes ago, dada051 said:
I never had problems with BRTFS, and I think a lot of users either. But why not. If you want to do VM backups, use can use QCOW2 as disk images, and https://github.com/wbynum/QEMUBackup as a docker to create backups.
This is just an rsync job to copy the vm. Non incremental and "best" when VM is shut down.
I get a lot of comments, that BTRFS is reliable - hasn't ever been for me. Sure, if you barley do something with it, it seems stable. Once you put some pressure on the FS is starts degrading pretty fast.
-
Another feature I would love it the possibility to include custom notification sinks via a simple webhook.
-
I voted for qemu, but frankly speaking I don't want any of those features. If I had a wish it would be the same as many other users have; snapshots/vm backups. I am running 4 productive servers and two lab ones. Here is why I think that there is a HUGE gap between what is supposed to work and how it actually works;
BTRFS: Ever since, the only "bad" experience I ever had, was related to BTRFS - no matter how I did set it up, it always failed at some point. I was running hourly checks, regular balances, etc. Still it kept failing, so I moved away to XFS. Therefore snapshots on BTRFS are not an option.
ZFS: I'm not a huge fan of a Filesystem, that eats tons of memory. For me, memory is my bottleneck for virtualization as it isn't shareable across VMs and/or other services. Due to my needs I use HEDT hardware over server stuff for the VM boxes and memory is limited to 128 or 256gb there. So that one is out as well.
VM Backup Script/Plugin: That thing is nice, but it does require the VM to be shutdown and will always take a full backup. My VMs (and for a lot of people out there) are running 24/7. Some are servers, others are just a pain in butt to start (hello to all the developers out there!). Also being a full backup, it takes extremely long to create, transfer, etc. Not talking about the extra disks it eats. Storing about 30 days per vm is somewhere around 1~3TB. That's a ton of space!
QEMU Tools: virtbackup isn't stable yet and not integrated in unraid. There are also some modifications required on the VM which tend to be overwritten (sometimes) by editing the VM via the UI or a server reboot. Also The interaction is not very comfortable (you can't remove or restore VMs via the UI). Rememer, you also need some docker to run a distro which would deliver these tools to your server.
As for now, I am working on a backup/snapshot solution based on a QEMU docker, but there is a million things which do require attention. Once It's done, I'll be providing it to the community, but I'd say this is a feature which should be included in unraid.-
2
-
-
Can I create and add a plugin for a custom notification sink?
-
We have several polls running about the specs of our servers, but what about the amount?
I'm really interested how many servers single users are running, could we get a poll for that?
-
Hi Guys,
I'm wondering if it is possible to add some custom notification agent. My issue with all the existing once is the lack of support for mqtt (or any reliable messaging architecture) and therefore a loss of messages in several use-cases.
Ideas`? -
WTF?! I tried to fix the btrfs issues and ran into this...
Please do NOT say "memtest" - the memory is fine, I tested it.
If I do boot in save mode, I don't get any panic... how can I track the issue down?!
-
-
Hi Guys,
My BTRFS pool just went read-only.
Aug 10 22:28:49 Vortex kernel: BTRFS error (device nvme0n1p1): bdev /dev/nvme1n1p1 errs: wr 12, rd 11354, flush 1, corrupt 1, gen 0 Aug 10 22:28:49 Vortex kernel: BTRFS error (device nvme0n1p1): bdev /dev/nvme1n1p1 errs: wr 12, rd 11355, flush 1, corrupt 1, gen 0 Aug 10 22:28:50 Vortex kernel: BTRFS error (device nvme0n1p1): bdev /dev/nvme1n1p1 errs: wr 12, rd 11356, flush 1, corrupt 1, gen 0 Aug 10 22:28:50 Vortex kernel: BTRFS error (device nvme0n1p1): bdev /dev/nvme1n1p1 errs: wr 12, rd 11357, flush 1, corrupt 1, gen 0 Aug 10 22:28:50 Vortex kernel: BTRFS error (device nvme0n1p1): bdev /dev/nvme1n1p1 errs: wr 12, rd 11358, flush 1, corrupt 1, gen 0 Aug 10 22:28:50 Vortex kernel: BTRFS error (device nvme0n1p1): bdev /dev/nvme1n1p1 errs: wr 12, rd 11359, flush 1, corrupt 1, gen 0 Aug 10 22:28:51 Vortex kernel: BTRFS error (device nvme0n1p1): bdev /dev/nvme1n1p1 errs: wr 12, rd 11360, flush 1, corrupt 1, gen 0 Aug 10 22:28:51 Vortex kernel: BTRFS error (device nvme0n1p1): bdev /dev/nvme1n1p1 errs: wr 12, rd 11361, flush 1, corrupt 1, gen 0 Aug 10 22:28:51 Vortex kernel: BTRFS error (device nvme0n1p1): bdev /dev/nvme1n1p1 errs: wr 12, rd 11362, flush 1, corrupt 1, gen 0 Aug 10 22:28:51 Vortex kernel: BTRFS error (device nvme0n1p1): bdev /dev/nvme1n1p1 errs: wr 12, rd 11363, flush 1, corrupt 1, gen 0Scrub does not seem to do anything.
a check -> btrfs dev stats -c /mnt/cache returns....
[/dev/nvme0n1p1].write_io_errs 0 [/dev/nvme0n1p1].read_io_errs 0 [/dev/nvme0n1p1].flush_io_errs 0 [/dev/nvme0n1p1].corruption_errs 0 [/dev/nvme0n1p1].generation_errs 0 [/dev/nvme1n1p1].write_io_errs 12 [/dev/nvme1n1p1].read_io_errs 12152 [/dev/nvme1n1p1].flush_io_errs 1 [/dev/nvme1n1p1].corruption_errs 1 [/dev/nvme1n1p1].generation_errs 0SOS please
-
Hi, welcome.
Let's clarify on a few on your assumptions fist:
In general you run an array (parity based) and some btrfs pools (raid based). Depending on the setup you have 1-2 Disks for the Parity array and whatever you define on raid level which can fail without effecting your data.
If a disk fails, you can replace it using hot-swap... Or you can also "preinstall" a disk into unraid, but not use it for any array. Once a disk fails, you could stop the array and replace it (inside the configuration, not physically!) with the spare. This operations takes less than a minute and unraid will begin to reconstruct the data to your spare disk. The reconstruction can take a while, but your data is AVAILABLE all the the time!As for Plex; I do use a similar setup and travel more often than I am close to my media. I do not struggle to stream anywhere in the world - if you do, it is probably due to your upload bandwidth on your server (usually ISP based). Having that in mind, it takes significantly more time to actually synchronize data compared to streaming a reduced (adopted to the available bandwidth, which plex does for you) just part of it...
An offsite backup IS a good idea, but it is not how/why you would address your actually topics.
-
How is unraid involved into this? What you are looking for some kind of bare metal issue. The "raw" draft would be having the controller and some disks in one bay. A bunch of disks in the other bay and connect those via some cables to another (or the same) controller inside the first bay.
-
🧡
this or set metrics
Thanks!
-
1
-
-
Any ideas? Anyone?
-
this already the case, but I still see eth2 being used
-
Hi Guys,
I have 3 NICs inside my server, but only one has internet access. I'd like to set this as the "default" for unraid. How would I achieve this?
-
Hi,
My Server seems to have 100% log - I can't post the diagnostics (won't donwload it).
but what I see in the log is
May 31 04:20:22 Natasha nginx: 2021/05/31 04:20:22 [crit] 7857#7857: ngx_slab_alloc() failed: no memory May 31 04:20:22 Natasha nginx: 2021/05/31 04:20:22 [error] 7857#7857: shpool alloc failed May 31 04:20:22 Natasha nginx: 2021/05/31 04:20:22 [error] 7857#7857: nchan: Out of shared memory while allocating message of size 9431. Increase nchan_max_reserved_memory. May 31 04:20:22 Natasha nginx: 2021/05/31 04:20:22 [error] 7857#7857: *80243 nchan: error publishing message (HTTP status code 500), client: unix:, server: , request: "POST /pub/disks?buffer_length=1 HTTP/1.1", host: "localhost" May 31 04:20:22 Natasha nginx: 2021/05/31 04:20:22 [error] 7857#7857: MEMSTORE:00: can't create shared message for channel /disks May 31 04:20:22 Natasha nginx: 2021/05/31 04:20:22 [alert] 10902#10902: worker process 7857 exited on signal 6 May 31 04:20:23 Natasha nginx: 2021/05/31 04:20:23 [crit] 7876#7876: ngx_slab_alloc() failed: no memory May 31 04:20:23 Natasha nginx: 2021/05/31 04:20:23 [error] 7876#7876: shpool alloc failed May 31 04:20:23 Natasha nginx: 2021/05/31 04:20:23 [error] 7876#7876: nchan: Out of shared memory while allocating message of size 9431. Increase nchan_max_reserved_memory. May 31 04:20:23 Natasha nginx: 2021/05/31 04:20:23 [error] 7876#7876: *80247 nchan: error publishing message (HTTP status code 500), client: unix:, server: , request: "POST /pub/disks?buffer_length=1 HTTP/1.1", host: "localhost" May 31 04:20:23 Natasha nginx: 2021/05/31 04:20:23 [error] 7876#7876: MEMSTORE:00: can't create shared message for channel /disks May 31 04:20:24 Natasha nginx: 2021/05/31 04:20:24 [crit] 7876#7876: ngx_slab_alloc() failed: no memory May 31 04:20:24 Natasha nginx: 2021/05/31 04:20:24 [error] 7876#7876: shpool alloc failed May 31 04:20:24 Natasha nginx: 2021/05/31 04:20:24 [error] 7876#7876: nchan: Out of shared memory while allocating message of size 9431. Increase nchan_max_reserved_memory. May 31 04:20:24 Natasha nginx: 2021/05/31 04:20:24 [error] 7876#7876: *80250 nchan: error publishing message (HTTP status code 500), client: unix:, server: , request: "POST /pub/disks?buffer_length=1 HTTP/1.1", host: "localhost" May 31 04:20:24 Natasha nginx: 2021/05/31 04:20:24 [error] 7876#7876: MEMSTORE:00: can't create shared message for channel /disks May 31 04:20:24 Natasha nginx: 2021/05/31 04:20:24 [alert] 10902#10902: worker process 7876 exited on signal 6
Can't even open a console
Any "simple" solution?
-
lol. That was easy. Thanks
-
Hi Guys,
I just started a new Server - my first Ryzen try - after assigning some disks I tried installing Krusader. It took very long (talking minutes). Any Container operation I try, basicly takes "forever" - it does complete however.
I attached my diags, maybe I need to set something up differntly for ryzen?
-
My issue seems not be the same, as the drived do spin down, but spin "magicly" up again after a short while... and spin down again...
Tried the SAS PlugIn, but hasn't changed much.
Shall I post in that thread you linked?
-
diags are at my first post. I'm using an adaptec (1100) controller for all the HDDs. No clue how to find "the criminal". Disks are getting pretty hot as they are spinning non-stop
-
Update: I tried turning off docker, all scripts, etc. Nothing changes. The drives keep spinning up and down. Not sure how to proceed from here...
-
There are no VM on the machine and not even a share on the pool atm.
Aggregate Licenses
in Pre-Sales Support
Posted
Hi Guys,
I do own 4 licenses and would like to get some more. But the current licenses are bound to different e-mails. Could I move the existing licenses to the same address and get the new once there?
aaaaand, is there a plan for "floating" licenses?