Score:-1

Server

Can a SATA hard disk be removed from bay w/o shutting down after being marked 'failed' in RAID1 array?

jamminjames

4/9/24, 7:55 PM

We have a software RAID1 array set up in a Supermicro 1u server. Can a SATA hard disk be removed (in a bay drive setup) w/o shutting down the machine? This would be after marking it 'failed' in the RAID1 array.

If so, I would assume you could replace it as well without shutting down, correct? Then you'd go through the process of making it part of the software array again.

Also, if that disk, once removed, were cloned, would that result in a bootable drive?

Thanks for any help.

1 + 0

clone

raid1

Score:0

Server

TomTom

4/9/24, 8:17 PM

You seriously are unaware that SATA connectors are safe to remove while operating? That was one of the reasons that the connections in them have variable length. Yes, you can safely remove it, if it is not in use (safe). It already failed - so it should be safe (check OS or whatever does the Raid 1) manual - but from the hardware level, it is absolutely ok to remove the HDD and replace it.

+ 11

jamminjames

4/9/24, 10:23 PM

Okay, thanks. And what about part 2 of my question. Would cloning the removed disk, having been part of a software RAID1 array, result in a bootable clone?

TomTom

4/10/24, 7:20 AM

There is no cloning unless you run really stupid software - the new disc should automatically be used to recreate the RAID 1, WORST ASE when you declare it a hot spare. Any hardware controller I know of would identify the new disc as replacement for a failed one and automatically integrate it and rebuild.

jamminjames

4/10/24, 7:33 PM

It's not a hardware array, it's software, as noted in my initial question. The intention is to use a copied disk for a separate server (as a backup machine), which is why I am inquiring about cloning it, sorry I wasn't clear on that. After making a copy, I'd like to put the original back in and re-sync the array.

TomTom

4/10/24, 7:52 PM

And unless you run really stupid software, it should behave like hardware, or? Man, please try understanding. Also, try reading the manual - as you STILL do not BOTHER to tell us WHAT software you use, you essentially tell us to guess. A little professionalism, please.

jamminjames

4/11/24, 4:52 AM

Please don't be so short with me, TomTom, speaking of professionalism. I'm new to this stuff. I thought there was only one type of RAID software on Linux, mdadm, which is what we are using. I've read up on it, but can't find the answers I'm looking for.

TomTom

4/11/24, 9:14 AM

"I'm new to this stuff." - this is a place for professionals. Read the site rules. "I thought there was only one type of RAID software on Linux," - and you did not THINK at all, right? Because you ASSUME that we KNOW that you run Linux. Your post is neither tagged with Linux, nor does it mention Linux (or mdadm) in the whole text. There are still other operating systems, among them i.e. Windows. You may have heard of that. And NOT being a Linux user, I found the relevant part in like 10 econds using Google and - and that even made me a small checklist - bing (the new AI hat version).

TomTom

4/11/24, 9:15 AM

The last one is - interesting. This is now a QUOTE...

TomTom

4/11/24, 9:15 AM

According to a source 1, you can rebuild a RAID 1 array after swapping out a defective hard disk using mdadm. First, you need to set the faulty partition using mdadm --manage --set-faulty /dev/mdN /dev/sdX1. Then, you can check the status of the software RAID using mdadm -D /dev/mdN. After that, you can hot remove the faulty disk using mdadm --manage /dev/mdN -r /dev/sdX1 and then add the new disk using mdadm --manage /dev/mdN -a /dev/sdX1. The rebuild process should start automatically.

TomTom

4/11/24, 9:17 AM

And the question was: "mdadm - how do I rebuild a raid 1 after swapping out a defective hard disc?". Source 1 points to https://www.thomas-krenn.com/en/wiki/Mdadm_recovery_and_resync

jamminjames

4/11/24, 7:28 PM

I can't find anywhere where it says Server Fault is only for experts. It says, "Server Fault is for questions about managing information technology systems in a business environment." That's what I'm doing, we have a business that uses a dedicated server. It also says: Always be polite and have fun. It’s fine to disagree and express concern, but please be civil. There’s always a real human being on the other end of that network connection, however misguided they may appear to be. That's all I'm asking for. Thanks for taking the time to answer, in your own gruff way, though, TomTom

jamminjames

4/11/24, 7:59 PM

Btw, yes, I have found many examples similar to the one you linked to. Some said you need to shutdown the machine before putting in a new disk, so there was some confusion there I was trying to clarify. Also, I can't find anything to answer part 2 of my question, regarding cloning the removed disk, and whether the clone would be bootable. Your link does not address that either.

Elon Musk

I sit in a Tesla and translated this thread with Ai:

EN: Can a SATA hard disk be removed from bay w/o shutting down after being marked 'failed' in RAID1 array?

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.