Beiträge von auanasgheps

Introduction

My NAS experience is based on the last 5 years using 2 different Synology NAS, but it’s been a while I wanted something DIY.

I optioned OMV two years ago but my Linux knowledge wasn’t there yet, so I bought my 2^nd Synology, the DS918+ - the high-end home NAS from Synology.

Today I’ve finally decommissioned this little guy, which was/is pretty expensive. I’ve paid a hefty 550€ at the time and the price hasn’t changed much since then, so my goal was to replace it with something cheaper.

This project would have not been possible without my friend Lorenzo who helped me a lot especially with SnapRAID concepts and the build itself.

Hardware

The goal I set was fairly easy: better hardware at a lower cost (DS918+ Specs are here). I added some caveats: no server parts (unlike other “NAS killers” out there), low power consumption when idle, recent hardware and reasonable silence.

To add some context, my server is mainly used for backup, downloads and storing my data. There’s only me using it for the most time, and a couple of friends use it for streaming over Plex.

These are the parts, with a little justification.

• CPU:Intel Core i3-8300 3.7 GHz Quad-Core Processor(Purchased For €63.00) – The only used component of the build. i3-8xxx series is basically i3-9xxx series and the price was great, so I went for it.
• CPU Cooler:ARCTIC Alpine 12 CPU Cooler(Purchased For €7.73) Cheap & silent, I also do not allow Intel stock coolers, anyone who does should be punished.
• Motherboard:ASRock H370M-ITX/ac Mini ITX LGA1151 Motherboard(Purchased For €98.50) - ITX but 6 SATA ports, no stupid RGB stuff, 2 onboard LANs, many USB. Approved
• Memory:Crucial Ballistix Sport LT 8 GB (1 x 8 GB) DDR4-3000 Memory(Purchased For €38.03) - The cheapest and low latency RAM I could find at the time of purchase: CL15 15-16-16. Running @2400 to save power.
• Storage:Silicon Power A80 256 GB M.2-2280 NVME Solid State Drive(Purchased For €45.00) - One of the few cheap NVMe SSD that has been reviewed by reputable websites and has great performance.
• Case:Fractal Design Node 304 Mini ITX Tower Case(Purchased For €84.03) Do I have to say something about this wonderful little case?
• Power Supply: be quiet! Pure Power 11 350 W 80+ Bronze Certified ATX Power Supply(Purchased For €49.10) Many SATA ports, reputable brand, silent.

Total: €385.39 -Prices include shipping

Add 11€ for a Sandisk Ultra Flair 32GB – PCPartPicker doesn’t allow for such parts.

All things considered, I made it under 400€ and this is great in my opinion.

Software

I’ve configured the system as follows:

• Sandisk USB drive – OMV boot device
• NVMe SSD drive – Docker and downloads storage
• 3x4TB 3.5’’ HDD drives – SnapRAID

Drives are coming from the previous NAS, and yes I know I have very little storage: I have a long way to become a data hoarder!

I’ve been testing OMV to become familiar with its concepts for a few months in VMs (only OMV 5), writing in this forum and digging a lot to document everything I could. This work turned out in an 8-page manual that has become my little treasure. For the record, the majority of the most useful things have been written by ryecoaaron, lol.

It’s been tough, I also reported some bugs during this period of testing, but after some hiccups I finally made it.

I use SMB for mounting shares to my 2 Windows computers. This one is the only integrated service I use; all the rest are plugins from omv-extras and dockers.

The system is very silent even though I have a discrete number of fans (3 case fans, CPU fan and PSU fan). It consumes 19W when idling with all disks in spindown (I hope somebody isn’t screaming for this).

Plugins I use: openmediavault-flashmemory, openmediavault-snapraid, openmediavault-unionfilesystems, openmediavault-backup, openmediavault-diskstats, openmediavault-resetperms, openmediavault-wakealarm, openmediavault-usbbackup.

Dockers I use: Dashmachine (great replacement for Heimdall), Glances, JDownloader, Krusader, let’s-encrypt/nginx, Plex, Portainer, Syncthing, Tautulli, Transmission VPN, Watchtower.

I still have to configure NextCloud (the docker compose it’s ready though) which I will use for accessing data from the web and for WebDAV features.

I can’t consider myself an OMV expert but I think I’ve accomplished a great setup for my needs and rounded a lot of rough edges. I am experienced at Docker, though. Don’t hesitate to ask questions about OMV or Docker configurations!

I have learnt a lot from this experience. I’ll try to publish the guide/KB I’ve written (it’s not in English and is rough in many aspects). I wanted to thank again the forum in general, there is a lot of knowledge here!

As a bonus pic, here’s my NAS killer next to his victim, for one last time. I won’t show you the internals: it’s messy.

Bingo, so you've found the issue! I'm glad my incident has helped.

As a side note, is there a way to customize the virtual filesystem name created by the plugin? It's ugly but most importantly long, I have to copy and paste it in every configuration (such as Docker files), would be handy if the user could customize it.

Zitat von ryecoaaron

None. Just trying to help. Wasn't sure if you knew what you put in the fields and/or had cut&paste the fstab to compare differences.

Thanks for you effort, always appreciated. I understand at this point we can only make speculations. Unfortunately I didn't know I had to check fstab otherwise I would have saved such file.

Well, below you can find my current fstab (working). Things to note:
- Made changes as instructed by the flashmemory plugin (updated root partition properties and disabled swap partition)
- Removed "noexec" property from the nvme partition which runs Docker

The bind I tried to made with mergefsfolder was slightly different. I entered the following paths, because I wanted the exact folders to be merged.I added them as the GUI instructed, one line per folder
/srv/dev-disk-by-label-HDD2/HDD2/data
/srv/dev-disk-by-label-HDD2/HDD3/media
I cant' remember if there was a slash / at the end of the folders. Could its presence/absence have caused the problem?

# /etc/fstab: static file system information.
#
# Use 'blkid' to print the universally unique identifier for a
# device; this may be used with UUID= as a more robust way to name devices
# that works even if disks are added and removed. See fstab(5).
#
# <file system> <mount point>   <type>  <options>       <dump>  <pass>
# / was on /dev/sdd1 during installation
UUID=05e735e2-df32-45df-996b-1d3ba752a19c / ext4 noatime,nodiratime,errors=remount-ro 0 1
# swap was on /dev/sdd5 during installation
#UUID=1b60aa3a-7b3d-4812-bea5-c687f368537e none            swap    sw              0       0
# >>> [openmediavault]
/dev/disk/by-label/HDD3		/srv/dev-disk-by-label-HDD3	ext4	defaults,nofail,user_xattr,noexec,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0,acl	0 2
/dev/disk/by-label/HDD2		/srv/dev-disk-by-label-HDD2	ext4	defaults,nofail,user_xattr,noexec,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0,acl	0 2
/dev/disk/by-label/nvme		/srv/dev-disk-by-label-nvme	ext4	defaults,nofail,user_xattr,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0,acl	0 2
/dev/disk/by-label/HDD1		/srv/dev-disk-by-label-HDD1	ext4	defaults,nofail,user_xattr,noexec,usrjquota=aquota.user,grpjquota=aquota.group,jqfmt=vfsv0,acl	0 2
/srv/dev-disk-by-label-HDD2:/srv/dev-disk-by-label-HDD3		/srv/fb4ab0da-2b79-4740-834a-7435c0123e28	fuse.mergerfs	defaults,allow_other,cache.files=off,use_ino,category.create=mfs,minfreespace=200G,x-systemd.requires=/srv/dev-disk-by-label-HDD2,x-systemd.requires=/srv/dev-disk-by-label-HDD3	0 0
# <<< [openmediavault]

Zitat von ryecoaaron

You don't want to post your setup and fstab?

Sure I can, but what is the point of doing it now when the system is up and running and using the "older" plugin?

Zitat von ryecoaaron

omv-extras 5.2.5 fixes the systemrescuecd ssh issue. sshd was running but systemrescuecd 6.0.4 added a firewall. I added nofirewall to the boot arguments and it works as expected. Just click install systemrescuecd again.

Thanks for updating systemrescuedcd. Probably you should not update that component, it's just a rescue CD
Regarding the mergefolders issue, I had the feeling this wasn't enough.
If I will play with it in the future I will let you know.

Whilst I was setting up my new OMV system, I wanted to give a try to openmediavault-mergerfsfolder, instead of using the "old" openmediavault-unionfilesystems.
Well, it's been a nightmare: it crippled my OMV installation twice before I realised it.

Disclaimer: I don't have many details and I don't want to replicate the issue because I already wasted a lot of time, but I thought was still worth reporting the issue.
OMV: 5.3.3-1
openmediavault-mergerfsfolder: 5.0.1

I installed OMV along with many plugins: backup, usbbackup, flashmemory, snapraid, resetperms, diskstats, wakealarm and mergerfsfolder.
I didn't configure most of them and didn't make any weird change to the system.

I configured mergerfsfolder to merge two main folders from different drives. Yes, I used /srv/dev-by-disk-by-label-XXX .
It seemed to work fine, I could see the new filestystem, browse it via WinSCP, mount it in SMB and do all sort of things.
Until I rebooted the system.

• I could not login to OMV UI - I think the error was very similar to this thread
• Any command issued via shell such as omv-update or other simple tasks resulted in "Read Only File System"
• I was able to remount the root FS and login to WebUI issuing the command mount -o remount / but the issue was still there If I rebooted again, services like SMB would not start (RED status in Dashboard)

I thought was my fault. Before rebooting I also updated from 5.3.2-1 to 5.3.3-1, so I thought this could have caused the issue.

I had to start all over again, because I hadn't made a backup yet ( ),
This time I configured everything but mergerfsfolder, and made a backup first. Rebooted, everything went fine, so the backup was healthy.
I configured mergerfsfolder and as before I could use correctly the merged FS.
Rebooted and bang! The same issue.

I managed to remount the root fs via the command above, install systemrescuecd and boot to it. By the way, it does not start ssh automatically (nothing was listening on port 22 but the system was connected correctly as I could ping it)

I restored the backup, removed mergerfsfolder and installed unionfilesystems.

I configured unionfilesystems, rebooted and everything has been working great since then.

Sorry for the rant, I hope this will help.

Zitat von votdev

As said and shown, postfix is already configured to handle such situations. The question now is why it does not work.

Hey, just wanted to let you know the issue does not occur any more.
Cron jobs are now received with the sender Cron Daemon, so I think it is working correctly now.

After reporting the issue here, I managed to replicate the issue in my following test installs, but it's been a while I haven't seen the issue.

Zitat von aoprea

This is drawback of omv vs synology or qnap... Why i so hard to have a file manager on omv ??? This method (mc over ssh conection or another method ... ) request that on every pc you try to connect to omv to have or install putty or winscp....

Too bad

Sent from my GM1913 using Tapatalk

Use Krusader in Docker as suggested by @KM0201. I've quickly tested and works quite well. You have the complete GUI as it uses VNC.
It's the most complete I've tested so far.
(I don't want to use SSH over the internet)

Thanks, will try your suggestions!

I think I've found a good File Manager:
https://filerun.com/

It can be installed via Docker, obviously.
Do you have other options?

Hi all,

Coming from a Synology NAS, there is one thing I still haven't found a replacement for: their web based File Manager called "File Station".

I am going to use NextCloud for general usage and sharing, but it's not so focused on file management tasks like "File Station" or a real file manager is.

As an example, Nextcloud by default doesn't extract archives, and even with the plugin enabled its support is somewhat flaky.

Do you have any suggestions or tackle the issue in a different way? I got used to a such powerful tool In a web browser!

Seems that Votdev has fixed the issue with this commit
https://github.com/openmediavault/openmediavault/pull/616

Yeah

Zitat von pvnnpavan

min receivefile size = 16384
write cache size = 524288
getwd cache = yes
socket options = TCP_NODELAY IPTOS_LOWDELAY SO_RCVBUF=65536 SO_SNDBUF=65536
log level = 1
read raw = yes
write raw = yes
max xmit = 65535
dead time = 15

Alles anzeigen

I am jumping in this conversation because I would like to know where did you get these extra options and if they are actually useful. So far I've only enforced SMB2 and above.

EDIT: I don't have technical experience with these parameters, but after few readings seems like that most of them are unnecessary or debatable. What is the performance without them?

Zitat von votdev

The issue must be fixed upstream (Debian or smartmontools). The OMV project can't do anything here. SImply create a filter rule in your email client that redirects those emails into trash in the meanwhile.

I already updated smartmontols to the actual latest version for another issue - wasn't detecting the NVMe drive for SMART reads, so it has something to do with Debian.

Zitat von ryecoaaron

Personally, I don't think you need smart for nvme especially if you leave some unpartitioned space on the drive. Most nvme drives are rated to write their entire capacity at least once a day for at least a few years. You should never hit this unless you are doing something crazy with it. If you just check the life left every six months after about 3 years, you would be probably know when it is going to need to be replaced.

Thanks for this advice, I will leave some unpartitioned space on the drive, disable SMART reading and rely on the hdsentinel tool to send me scheduled emails.

Hi,
OMV5 (5.3.2-1) is logging a "useless" NVMe property that leads to scary SMART alert emails that make me worry for no reason.
Here is what I am talking about.

This message was generated by the smartd daemon running on:


   host name:  [blip]
   DNS domain: [blop]


The following warning/error was logged by the smartd daemon:


Device: /dev/disk/by-id/nvme-SPCC_M.2_PCIe_SSD_SerialNumberRemoved, number of Error Log entries increased from 40 to 42


Device info:
SPCC M.2 PCIe SSD, S/N:SerialNumberRemoved, FW:ECFM13.1, NSID:1, 256 GB


For details see host's SYSLOG.

This drive is new, and I've checked its health with other tools, and obviously it's doing great.

The report is referring to Error Information Log Entries. A quick search reveals it is not due to the drive health.

Some interesting discussion here from bugs.debian.org.

I am no expert of this tool but does somebody know what could be done, like tellling the utility to ignore this value (maybe it's not NVMe aware)?
Honestly I want SMART enabled for this drive, but only for real warnings.

Hi,
I just installed OMV5 on my shiny new build and I have spotted this error coming up frequently in logs:

systemd[1]: /lib/systemd/system/nmbd.service:9: PIDFile= references path below legacy directory /var/run/, updating /var/run/samba/nmbd.pid → /run/samba/nmbd.pid; please update the unit file accordingly.

It's so frequent that appears in every log page.
- Samba is not (yet) enabled
- The installation is pretty clean, I haven't messed with any system file (yet)
- I am running OMV 5.3.2-1

I believe it's just a warning but it's very annoying. I have found some reference here, but should this be fixed?

Zitat von ryecoaaron

Debian not OMV but 7.1-1 is in backports but not pinned. It will be installed if you install proxmox kernel as well since 7.1 is in the proxmox repo. You can install from buster-backports with: apt-get install -t buster-backports smartmontools=7.1-1~bpo10+1

Thanks for the advice, will try tonight.

Since OMV got the 5.4 kernel officially, I thought there was no need for proxmox kernel. Do you still recommend it? I see proxmox it's 5.3 though

Zitat von Shinobi

it works in the shell with this command

smartctl -x /dev/nvme0

This specific SMART issue is because smartmontools is not updated in OMV5. The bug has been solved by smartmontools team, but the updated version is not showing up. Found it out just a while ago.
NVMe SMART status reading failure (wrong drive name)

Hi,
I am sorry to reply to an old thread, but the issue still exists in OMV5, I am currently running 5.3.2.1 and have the exact same issue as OP:

• nvme drive shows up as nvme0n1
• SMART data is shown correctly if I run via shell smartctl -x /dev/nvme0
• Error message that in the webGUI "Read NVMe SMART/Health Information failed: NVMe Status 0x2002"
• If I run smartctl -x /dev/nvme0 I get some data and the same error as above

smartmontools version is 6.6.-1. The issue has been fixed by the smartmontools team, but OMV does not pick up the updated package.
This is a newly installed system.

That's a great idea! Please make it for OMV 5 too!