JonathanM

@SpencerJ

Most times this is for speed reasons, not required. If memtest runs ok with 3 dimms, Unraid should as well.

I recommend searching for Spaceinvader One's youtube videos on Unraid, he has a great set of videos describing how Unraid works and how to implement all sorts of things.

School of hard knocks. Most times it's the knowledge that you earn the hard way that sticks with you best. 🙂 BTW, I don't remember if it was you or not, don't think so, but someone in the last day or two made a comment that their memory was new, so it shouldn't be bad. Given the lack of quality control in computer parts, the only memory you can assume is good is what you personally tested. New memory, or any time you make RAM changes, you should run a memtest cycle. Too important to skip.

I've personally had a motherboard that would throw memory errors only when all 4 slots were populated, so it's necessary to test the system as it will be run. In your case, you seem to have found a genuinely bad stick, hopefully you can replace it and be good. In my case, memtest came up clean on any given pair, in any given pair of slots, but once all 4 were together, it would error every time. You need to get at least 1 full clean pass with the memory exactly as you plan to run the system. The downtime is a necessary evil to ensure the best chance of future success.

Which means all the time spent rebuilding disks was wasted, as you can't trust what was written. All the bits and bytes that are read from and written to the disks interact with RAM, so those errors very well could have corrupted any data written to the hard drives at any point in the process. Never use a computer with suspect memory to work with files you care about.

No. UEFI is fine, but you must use a different USB stick to run memtest if you use UEFI.

I don't remember the magic sauce, but if you format them in UD, you should be able to remove them and add data in any other distro that understands XFS. All root folders on all the drives will become user shares with their content merged in the /mnt/user/folder structure, so you can set up shares at the same time by judiciously using root folders.

That's an old version, and Pro is not available for free. I recommend downloading the current version directly from the legitimate site, not an old version from some questionable source.

memtest

I don't understand the question. The attached picture shows /dev/sg0.

Neither did I, except in the logs as an error message. I was watching the collabora container logs in realtime as I attempted to edit a document, and it called out my FQDN and said it wasn't authorized in the alias_groups. So I added 1+1 and got 3. If that is indeed the case, and all the stuff I found online with aliasgroup1 is erroneous, I'm going to be rather upset that the internet wasted half my day. 🤣 All I know is, I made the change, and now it works, and no nastygrams in the logs about my domain not being authorized.

I just had a breakthrough, don't know if it's luck or the actual solution, but for me, Key: alias_group1 Value: https://my\\.nextcloud\\.domain:443 worked, finally. Apparently the _ between alias and group1 is quite important. Don't know how others have gotten it to work, but that's what did it for me.

Skip the preclear, just replace the drive, build parity, do a parity check, do a long smart test on the drive. Preclear is not productive for your situation, you want to get valid parity back in place ASAP. If you had a new drive in anticipation of replacing a drive that was still functioning, then preclear might be of some value, but only as a stress test to weed out infant mortality. You will get that same level of testing by rebuilding the drive and checking parity, and the bonus is you will be protected that much sooner. If the new drive fails during the process, you aren't in any worse shape, just back to needing another drive. If you were getting a new drive to ADD to a NEW data slot, not replacing a drive, then preclear could save some time as well as give you a confidence check in the drive before trusting it in the array. This does not apply to your situation.

I was just going by the manual that shows 2 cables that come with it. https://download.schneider-electric.com/files?p_File_Name=SU+QS+990-5460C+MN01+EN.pdf&p_Doc_Ref=SPD_AHUG-A5Y9GY_EN&p_enDocType=Quick+start+guide Oh, I'm positive it's a serial connection, but APC is notorious for not using industry standard pinouts to force you to buy their specific cables, or make your own.

Unfortunately I don't think you will get any response. It's been almost 5 years since he signed in.

Oh, at some point there needs to be a section for the new file manager plugin if there isn't already, I didn't look for it. When that happens, the warning about mixing user and disk shares can link to the file manager since that should be a safe way to deal with the types of issues that lure people into mixing user and disk shares.

Content seems fine, proofreading could use some help 🙂 I also did some minor editing of the disk share section just now. Please review and fix if needed.

Have you tried the USB or serial cable that came with it? AFAIK the RJ45 "serial" port uses a special circuit that generic cables won't work.

Since you are playing with the wording, I have a couple nits. When taken in the context of the full section this seems to imply that the disk shares won't be available under /mnt unless disk shares are enabled. I don't have a suggestion on how to reword this, because it's factually accurate as it stands, but I can see how someone could be misled into thinking they needed to enable disk shares to see the /mnt/diskX paths.

Was the appdata backup running? It shuts down the containers then restarts them when the backup is complete.

Probably. However, I was able to get the container running by removing the domains value in the template. I haven't got it connected back to nc yet, but apparently there is a way to do it, I just haven't had time to work on it.

It runs a half dozen containers, along with a Win10 VM that seldom gets run, it's only a backup. The main load is a mining container and UrBackup with about 8 active clients. Does everything I ask of it, until you commented I hadn't even thought about how old the platform is. It's amazing what Unraid can get out of old hardware.

@dlandon, could this be used instead of a ping for smb connections?