JonathanM
-
Joined
-
Last visited
Posts posted by JonathanM
-
-
-
When you no longer trust it. All remaining disks must be read flawlessly to reconstruct a failed disk, so if another disk you don't suspect of failure fails, then you are relying on a disk you don't trust to rebuild it.
As far as how to tell? That's part intuition and part analysis. All disks will fail eventually given time, the gamble is waiting until the last moment. Increased re-allocated sectors that happen recently and the count continues to increase, that disk is very likely to die soon. Nobody can predict the future perfectly, so you just have to use your best judgement, analyze the SMART numbers for that specific manufacturer.
Some people have a very low tolerance for risk, and replace drives as soon as they show any sign of trouble. Some people like to live dangerously, and wait for a drive to actually get dropped from the array to replace it.
In any case, drive redundancy and the ability to replace a failed disk does not mean you don't need backups of anything you don't want to lose. There are MANY ways to lose data that don't involve hard drive failure.
-
1
-
-
On 1/13/2025 at 6:06 PM, greenskye said:
I almost always buy drives one at a time
That's arguably the safest way to do it. It eliminates a bunch of shared variables that can kill a drive prematurely, like shipping damage.
The last thing you want is to have a bunch of drives that were shipped with inadequate packing that FedUPS drop kicked across their warehouse. Or a whole box of drives that got dropped at the distributor, so they don't have any external signs of damage, just high g loads that probable won't matter, but could.
If you really want to get a bunch of the same drives, try to stagger the purchases ideally both in time and location. Never buy 2 that will ship in the same box.
-
-
-
-
On 4/18/2025 at 5:47 PM, sbm7815 said:
Would you mind sharing where were you able to purchase this board? I've been looking all over for it. It's always out of stock or not carried at all because it's so new. I'm thinking of just placing a backorder with Tech America.
My distributor in the USA says they will have stock on April 30th.
-
-
4 hours ago, doug-kasper said:
the `/etc/exports` on the Unraid server
That file only exists in RAM while the server is running, and is extracted from the archives on the USB every reboot. Any changes you make need to be redone after each boot. Best way is "CA User Scripts" plugin, write a short script to accomplish the edit, and set it to run "At startup of array"
-
-
-
Unraid doesn't support hot-swap.
The Unraid parity array does indeed give you fault tolerance for either a single or double drive failure depending on whether you assign 1 or 2 parity disks. It is NOT RAID5 or RAID6, the data is NOT striped across devices, each drive has its own valid filesystem. The parity drives are independent from the data drives and hold no sensible data, only the parity bits required to rebuild a failed drive.
The advantages of the Unraid parity array is greater resilience and possible power saving, as unused drives can be spun down, also if there are drive failures beyond the tolerance level each remaining data drive remains readable.
The disadvantages compared to RAID is speed, as reads are limited to a single device, and writes must update 2 or 3 devices. Also, since each drive is a separate filesystem, any single file can only be as large as the free space on any individual drive.
The filesystems from each drive are usable as a group via user shares, which can span multiple drives, each unique root folder on the data drives is a user share, and the file tree view of all those folders is merged when looking at the user share.
I suggest looking at some of spaceinvader one's videos on youtube, he has some good basics of Unraid as well as more advanced use cases.
-
-
-
-
-
10 hours ago, TallMan206 said:
Can you guide me into the right direction?
https://github.com/binhex/documentation/blob/master/docker/faq/help.md#gather-supervisor-log
-
3 hours ago, Rocka374 said:
Its Ubuntu VM. So the important part is just to stop the VM 1st. I more question. If the VM is completely new config with new MAC and I paste this vdisk there, it won't work right? Later if I configure my router to assign the old local IP to the new MAC, should it work again? Or everything in the config needs to be the same for the backup to work?
Depends on your definition of "work". If the MAC changes, it would be the same as if you swapped a NIC in a physical box. It would still boot, but your router would see it as a new device.
-
5 hours ago, Rocka374 said:
Dear, I see the share via SMB on my PC. I have a lot of space there. Is it possible to just copy the VM vdisk, manually? Will the VM work if I just paste that this later? Or if I configure new VM with the same IP and parametars?
If the VM is shut down when you do the copy, yes to all with some caveats. Setting up a new VM will assign a different hardware ID, so Microsoft Windows will assume it's a new pc. Linux VM's are much more forgiving.
-
-
-
27 minutes ago, Wody said:
If there was, they wouldn't supply that cable in the first place.
In principle I agree, but also this exists. I'm not disagreeing with splitting the molex 4 pin connector to multiple outlets, but the existence of this abomination shows that ability to source a cable isn't proof the cable is proper for the application.
-
4 hours ago, karateo said:
Is there anything that I am missing?
The value of the time needed to code vs. the actual value of the utility. It would only be useful in an extremely small number of cases, and all of those cases would be covered by a proper backup, which would be needed anyway for a much larger fraction of data loss episodes.
It's just not worth it right now to spend limited time (money) coding something where a solution already exists. If it is worth your time to work on coding it, nobody is saying to not do it.
-
1
-
-
On 3/6/2025 at 2:49 AM, Amiga1000 said:
I'm pretty sure the path exists
It doesn't, unless you map it.
Each container is an independent little linux install, it only sees the folders inside the container until you tell it how to see external paths. Look at the install template, on the path shown as Source Data on the container side, you put the /mnt/user path on the host side. Then when you look in the Source Data folder inside the container, it shows the contents of whatever folder you mapped there.
Parity is Already Valid - Should I Check?
in General Support
Format doesn't clear the drive, it just writes a new table of contents with no entries. All the old ones and zeroes from the deleted data are still there. To remove a drive without effecting parity, you must write all zeroes to the drive, which removes any existing format as well as any traces of previous data. Writing all zeroes takes a rather long time, typically many hours. It's normally much quicker to redo parity.