Questions about NAS

[Mystique]

@cobito One good thing about xpenology is that the usb is just the loader. The system is on the hard drives (it forces you to choose a volume during installation and puts itself on it)

One thing I liked about FreeNAS (which I don't know if you would eventually run into the same problem) is that the system could be installed on a couple of usb (basically, a raid 1 of usb for the system), so you were relatively protected from these cases

But well, everything can break. As a security measure, an image of the drive on your PC to recover it if it breaks again is not a bad idea

[cobito]

I've been quite busy this past month and haven't been able to tinker with the NAS. In the end, I didn't try xpenology (although I had it in mind) and at this point, I'm going to stop experimenting.

There are a couple of things I've noticed. On the one hand, from the server, I get transfer rates of around 100Mbytes/s, just like from my Windows 10 PC (which I expected from the beginning). But for some reason, when I boot Linux on my PC, it stays stuck between 60-70MBytes per second, so I'm putting it down to something related to the PC, even though the CPU is far from reaching 100%.

Another thing is that when a file transfer fails, it gets stuck on the NAS, so you can't delete it via SMB. There are only two ways to delete it: access it from SSH as root and delete it or restart the NAS. When there are many files stuck, periodically, the NAS stops responding for a few minutes, which resets the connections of all the devices connected to it. If during this disconnection you were writing to a file, the file is left incomplete. This is dangerous for the server because you have to check the log periodically. To avoid this, you just have to restart the NAS if for some reason, the transfer of a file has failed. If the NAS is fresh, it can go weeks without producing this kind of blockage (and I say weeks because that's the time I've been able to test so far and I hope it's indefinite if no transfers are messed up).

Overall, I'm happy with the device. It's been a relatively important investment (equivalent to the components to build a pretty decent PC), but if the hardware doesn't give me problems in the coming years, it will have been worth it. On the one hand, I've been able to centralize the storage of several devices, which gives me convenience and security. On the other hand, I have a storage device with a large capacity and very good performance in random access, which I can expand and manage as I see fit, which ultimately, was what I was looking for.

[_Neptunno_]

Of course it's great to read your experiences and impressions about NAS, OVM and Xpenology. I have so far made moderate use of my little "gadget" with OMV and I will probably have to refresh myself a bit on the subject although it is true that before leaving it running I did make a USB image just in case, in addition to installing a Plugin (which, for the sake of it, I can't remember how I did it) to avoid the degradation that these devices suffer.

@cobito I will tell you that I used a 16GB 2.0 Toshiba from Amazon (about 8€) and so far it keeps working without any problem. But I think it's always good to have another pen usb ready in case the one you have fails. The only thing is if you make modifications and in the emergency one they are not implemented those modifications. Of course I'm talking to you in a simple and straightforward way, surely there are a thousand ways to have the other copy or image updated

Anyway, thanks to both of you for your advice and experiences!!

[cobito]

One year later...

With just a few days to go until the NAS turns one, one of the drives has failed. For the past few months, there had been an occasional clicking sound. A few days ago, the clicking became more frequent, so I witnessed the drive's demise in real time. The NAS spent a few minutes trying to access it, and during that time, the file system was inaccessible. After that, the failed drive's activity light stayed on permanently, and the RAID continued to function normally. That same night, I ordered a replacement drive (the truth is that the 4TB Red NAS drives from EFAX are becoming less common). I received it yesterday and ran a surface scan (as I always do when I buy a mechanical drive). Last night, when it passed without any issues, I proceeded with the replacement.

It has been rebuilding the RAID5 for 11 hours and is back to where it started, with the NAS in general and the RAID in particular performing as expected. Now I need to process the RMA for the failed drive, and I'll keep it as a spare or as an expansion for the RAID if I run out of space (whichever comes first).

Regarding the hardware, the motherboard with integrated CPU seems like a good choice. In addition to being quiet and low power, I haven't had any issues with crashes, instability, or anything else. It also offers more than enough processing power. Even during the RAID reconstruction, which is a costly operation, neither of the cores reached 100% (the rest of the cores were idle). The NAS has been running with the 2.5" USB external drive with the system since the USB stick failed, and although it's not ideal, I think it will continue like this. The best future solution will be to put an M2 SSD drive.

And about the case, it seems that the blue LEDs indicating the presence of a drive in the bay are turning off. They might be burning out. It could be dirt, but the green activity LED (which is right below) is as bright as it was on the first day, so it looks like it's a design flaw.

[cobito]

Two months later...

I have received a replacement drive. I don't know if it's normal or not, but the process seems to have been a bit slow and with little information. In case anyone is curious, the RMA process with Western Digital has been like this:

• On September 12, I created an RMA
• On the 16th, I made the shipment.
• On September 20, the drive was received (according to Correos) at the logistics center that WD has in Ontígola, Toledo.
• Seeing that the status did not change, after two weeks or so, I wrote to WD, at which point they updated the status to 'processing'.
• On October 18, they sent me an email saying that they had received the drive.
• On November 3, the RMA status changed to 'Completed pending return'.
• On November 11, I received an email from WD saying that my drive had been shipped on November 3 and was waiting to be picked up at a UPS collection point since November 4.
• Today, November 12, I went to pick up the drive.

Apparently, they were about to return it for exceeding the maximum time, but I arrived in time.

The good side: it looks like they have given me a brand new drive with a manufacturing date of August 31, 2021, so they are getting away with it.

Tonight I will do a surface scan and leave it stored for when I need it (expansion or replacement of a defective drive).

[jordiqui]

@_Neptunno_ I had two scassi in raid 0 and one for storage and I never had any problems with the raid. I think it was on Windows Server 2000.