c3

1) use another system, doesn't need to be much to just run preclear for 1,2 drives. 2) preclear on USB, much slower but often the drive comes as external.

The HGST NAS are about the highest rated drives.

Thanks for the XFS, as dev I can not really come here and promote it. 1% of 8TB is 80GB, your 120Gb is 1.5% which you probably did to avoid the warning at 99%. I :+1: your request for getting the setting to allow decimals as I am constantly dealing with people who claim out of space 100%, yet have 100s of GB available. When you have room for thousands/millions more of your average filesize, you're not out of space, just because you see a 100% from df.

It depends... If the data is written, not deleted, changed, or grown, you can fill a filesystem very full. However once you begin making changes, called making holes, or fragmenting, things can become very ugly. Depending on the filesystem, you may notice as soon as 80% full, but certainly in the 90+% range the filesystem begin doing a lot of extra work if files are being changed. Also, almost all filesystems keep working on this to improve. That said, you want to put 8TB of cold storage, (what about atime, etc?), you can go very full 99%, just be sure it is truly cold. I would even single thread the copy to avoid fragments, but that is really just a read speed thing.

The guidance is to use -90% or -95% in BES. What are you using?

The googlefu is not strong in this one... https://forum.videohelp.com/threads/369116-dvd-shrink-out-of-memory-error Your cache drive idea is reasonable since write performance seems to be the cause, or as suggested in the thread above, hit pause manually, allocate huge memory.

Last month RedHat also took a step back from btrfs From the 7.4 release notes.

I did not mean that all disk need to be the same size (or new), rather the typical new build has all new drives, and typically all of them are the same size. I think it is unlikely that someone would purchase drives of differing sizes for a new unRAID build, but it can certainly happen. Certainly, many have inserted various on-hand drives in the initial build. For example, if I was building a new unRAID, it would have all 8TB drives. As you asked, how much capacity is needed? let's say 20TB. Would you really buy (4) 8TB and (1) 4TB? or just (5) 8TB? That's (2) parity + (3) data, 8+8+4=20TB or 8+8+8=24TB And if 6TB is the deal of the day/week/month, (5) 6TB and (1) 2TB or just (6) 6TB? That's (2) parity + (4) data, 6+6+6+2=20TB or 6+6+6+6=24TB

On a new unRAID server (all new disks) parity size is equal to data size. ie, all 6TB or all 8TB. Later, a new 12TB comes out, and you get a black Friday deal, the 12TB goes in as parity before any data can be 12TB.

The one thing that does not matter is age. You should replace a drive giving errors. You can/should replace a drive if you need capacity. You can replace a drive for performance, but in most cases this does not apply to unRAID data drives. There is a thread(s) around which covers doing performance testing to find potentially "bad" drives. The idea is that by testing the performance the drive may show problem areas as dips in performance. If I run across the thread, I'll edit this post, or someone can add a link.

Agreed, hence I do use an external fan and some cardboard ductwork. I have not come across this issue. However the purpose of this preclear is to find DOA drives, which remains valid.

I just preclear before shucking. yes, it takes longer, but it typically finishes before I can get back to work on it again.

This is the topic of "preclear". You may want to do additional testing, but it is a good start. Most run just preclear some number of passes between 1 and 3. It does take some time, but worth it, to avoid the things you mentioned. I would not be concerned about ordering multiple drives at once. A more significant concern is when the drives are not properly packages for shipment. Buying from multiple sources will not avoid the historical "bad batch". They are more like bad model/size.

Did you win? How does BetterStreamingVideo.com, the next new innovation, get over the wall built by NetFlix and the cable company's back dealing? They created a barrier to entry. This is why you find that huge list of internet startups (and not so startup) opposite the ISPs. So, you counter, just switch ISPs, great thought. Yeah, they thought you might want to switch if you learned about it. ISPs like Comcast work directly with local law makers to write laws excluding other ISPs from providing competitive alternative to you. Remember Google Fiber was coming to Nashville, Comcast steps in, writes an ordinance with work rules stalling the competition. Barrier to entry, again. When forced to share they do as much damage as possible, tort action of Comcast against Telcom Cable. You are paying $45(plus fees)/month to have 1TB (their number) delivered to your home as quickly as possible, without delay or blocking. You did not sign up for the "chosen few of the internet" package. So, they don't get to chosen which data is delayed or blocked. You ordered internet access, access to all of the internet. Oh, this even works for free access, check out the mess Zuckerberg got into with free internet in India.

yes, and they sued, thus forcing the classification.

Just read paragraph 81, they want to remove the no-blocking rule by moving themselves back out of common carrier. Basically they got caught blocking, lied to congress, sued that was wrong to say blocking was not allowed, and so got put under heavy regulation where no-blocking rule is allowed. If they are allowed to return to old status, they will return to old ways. The only way to enforce was to change classification and they are whining about it. It is cat+mouse, keystone cops, everyone knows who the bad guy is, but he keeps slipping the cuffs and running. [and ditto for throttling, a lame way to say "we're not blocking" (just slowing to unusable)]

Similar experience with 5TB years ago. Not sure unRAID will ever use the f2fs, but once the kernel upgrades (4.10 and beyond), using the dm-zoned device will mitigate this further.

Why not just follow the procedure and upgrade the parity?

One at a time will not break anything. Two at once will be much quicker.

license is tied to usb guid, so while you can boot from hdd, the usb will be still needed.

In case anyone searches... https://arstechnica.com/security/2017/05/windows-xp-pcs-infected-by-wcry-can-be-decrypted-without-paying-ransom/ At least someone was able to decrypt without paying.

Were you running

A sector becomes pending when the drive is unable to read data from the sector. At that point data was irretrievable, or lost. As a pending sector, a write can be attempted, if successful, the sector is not reallocated. If the write is not successful, the sector is reallocated. Bad reads result in pending. Bad writes result in reallocated.

Previous thread has a lot of info

For drive 5 and 6 (or 3 and 4) At the end you will need to change the config and rebuild parity.