[PLUG] Raving mad RAID
About a week ago I finally was successful in creating a RAID0 array on my four NVMe drives that are installed in a Thunderbolt 3 enclosure. After creating the array it appeared in /dev as md0. After rebooting it became md127. I copied the UUID from Gparted and used it in a line that I added to /etc/fstab. The array has been working fine ever since I created it, including copying files to it late last night. This morning I tried to add a torrent for a distro ISO to Ktorrent, and got an error message that Ktorrent couldn't add the torrent because the location to copy it to did not exist. WTH? I looked at my GUI file manager and all the files in the array were listed. I right-clicked on one of them and immediately noticed that Rename and Delete were no longer listed in the options. After a bit more poking around I determined that the array had become read-only overnight. I decided to umount it and then re-mount it. The umount command gave me 'can't read superblock on /dev/md127p1,' which is what /dev/md0 became after rebooting a week ago. However, apparently the umount command succeeded, because it was no longer mounted. Then I tried to re-mount it and got the same superblock error message. Looking at /dev I see that most everything has changed. NVMe1-3 now have namespace 2 instead of the 1 that they were when I created the array. And now nvme5-8 are listed, which don't exist. And only nvme4n1 had a partition after I created the array, and now it has two partitions. It looks like I'm going to have to nuke the array, re-make it, and wait 24 hours to copy the 10TB of data back to the new array from the NAS backup. But before I do that I need to find out what went wrong. Might there be a defect in one of the NVMe drives? Or might there be a bug in mdadm when it tries to create an array out of NVMe media? Or when the ext4 filesystem was created? I assume that there exists a utility to check a drive, but I've never done that before. Suggestions? I'm considering throwing my computers into the river and doing something useful with my life. ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Mon, 2021-02-01 at 16:19 -0800, John Jason Jordan wrote: > About a week ago I finally was successful in creating a RAID0 array > on > my four NVMe drives that are installed in a Thunderbolt 3 enclosure. > After creating the array it appeared in /dev as md0. After rebooting > it > became md127. I copied the UUID from Gparted and used it in a line > that I added to /etc/fstab. > > The array has been working fine ever since I created it, including > copying files to it late last night. This morning I tried to add a > torrent for a distro ISO to Ktorrent, and got an error message that > Ktorrent couldn't add the torrent because the location to copy it to > did not exist. WTH? > > I looked at my GUI file manager and all the files in the array were > listed. I right-clicked on one of them and immediately noticed that > Rename and Delete were no longer listed in the options. After a bit > more poking around I determined that the array had become read-only > overnight. > > I decided to umount it and then re-mount it. The umount command gave > me > 'can't read superblock on /dev/md127p1,' which is what /dev/md0 > became > after rebooting a week ago. However, apparently the umount command > succeeded, because it was no longer mounted. Then I tried to re-mount > it and got the same superblock error message. > > Looking at /dev I see that most everything has changed. NVMe1-3 now > have namespace 2 instead of the 1 that they were when I created the > array. And now nvme5-8 are listed, which don't exist. And only > nvme4n1 > had a partition after I created the array, and now it has two > partitions. > > It looks like I'm going to have to nuke the array, re-make it, and > wait > 24 hours to copy the 10TB of data back to the new array from the NAS > backup. But before I do that I need to find out what went wrong. > Might > there be a defect in one of the NVMe drives? Or might there be a bug > in > mdadm when it tries to create an array out of NVMe media? Or when the > ext4 filesystem was created? I assume that there exists a utility to > check a drive, but I've never done that before. Suggestions? > > I'm considering throwing my computers into the river and doing > something useful with my life. > Perhaps now would be the time to dig out those old emails and consider some of the native alternatives rejected in favor of RAID0. Just saying, -T ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On 2/1/21 9:16 PM, TomasK wrote: On Mon, 2021-02-01 at 16:19 -0800, John Jason Jordan wrote: About a week ago I finally was successful in creating a RAID0 array on my four NVMe drives that are installed in a Thunderbolt 3 enclosure. After creating the array it appeared in /dev as md0. After rebooting it became md127. I copied the UUID from Gparted and used it in a line that I added to /etc/fstab. The array has been working fine ever since I created it, including copying files to it late last night. This morning I tried to add a torrent for a distro ISO to Ktorrent, and got an error message that Ktorrent couldn't add the torrent because the location to copy it to did not exist. WTH? I looked at my GUI file manager and all the files in the array were listed. I right-clicked on one of them and immediately noticed that Rename and Delete were no longer listed in the options. After a bit more poking around I determined that the array had become read-only overnight. I decided to umount it and then re-mount it. The umount command gave me 'can't read superblock on /dev/md127p1,' which is what /dev/md0 became after rebooting a week ago. However, apparently the umount command succeeded, because it was no longer mounted. Then I tried to re-mount it and got the same superblock error message. Looking at /dev I see that most everything has changed. NVMe1-3 now have namespace 2 instead of the 1 that they were when I created the array. And now nvme5-8 are listed, which don't exist. And only nvme4n1 had a partition after I created the array, and now it has two partitions. It looks like I'm going to have to nuke the array, re-make it, and wait 24 hours to copy the 10TB of data back to the new array from the NAS backup. But before I do that I need to find out what went wrong. Might there be a defect in one of the NVMe drives? Or might there be a bug in mdadm when it tries to create an array out of NVMe media? Or when the ext4 filesystem was created? I assume that there exists a utility to check a drive, but I've never done that before. Suggestions? I'm considering throwing my computers into the river and doing something useful with my life. Perhaps now would be the time to dig out those old emails and consider some of the native alternatives rejected in favor of RAID0. Just saying, -T Unfortunately it looks like RAID might not be the culprit if his NVMe /dev nodes are moving around. RAID0 isn't the cause but it will make things more complicated when something fails further down in the stack. If his system is dynamically naming devices in /dev/nvme* then that needs to be dealt with before even thinking about RAID. Not really sure where to start looking at that off the top of my head since I was under the assumption that this wasn't supposed to happen with NVMe. -Ben ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Mon, 1 Feb 2021 22:15:12 -0800 Ben Koenig dijo: >> Perhaps now would be the time to dig out those old emails and >> consider some of the native alternatives rejected in favor of RAID0. >Unfortunately it looks like RAID might not be the culprit if his NVMe >/dev nodes are moving around. RAID0 isn't the cause but it will make >things more complicated when something fails further down in the stack. > >If his system is dynamically naming devices in /dev/nvme* then that >needs to be dealt with before even thinking about RAID. Not really >sure where to start looking at that off the top of my head since I was >under the assumption that this wasn't supposed to happen with NVMe. There was recently a bit of discussion about LVM, and Rich sent me some links. I tried to read and understand it, but it seemed even more complicated than RAID. Plus, several years ago, when I was using Fedora, one of their obligatory updates changed my setup to LVM (without telling me that it was going to do so), and I couldn't get rid of it. That left a bad taste in my mouth and I have always avoided LVM ever since. But I must admit that my dislike of LVM is pure bias without much science. I am more concerned about devices renaming themselves and changing how they are mounted, all without any input from me. About January 20 I lost the first array that I had been running without a problem for about a month. And now my re-creation of that array is playing up after only a week. As I mentioned before, after rebooting the drives appear fine, read-write, but when I launched Ktorrent it complained that about half of the files it was seeding were missing. The files are all there and I can do anything I want to with them, but something is screwy with access. And why just half of the files? Either they should all work or they should all fail. Right now my money is on a defective drive. I need to find some tools for diagnosis and learn how to use them. All four are brand new Intel drives. If they all check out OK, then it's time to consider other possibilities. ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On 2/1/21 11:35 PM, John Jason Jordan wrote: On Mon, 1 Feb 2021 22:15:12 -0800 Ben Koenig dijo: Perhaps now would be the time to dig out those old emails and consider some of the native alternatives rejected in favor of RAID0. Unfortunately it looks like RAID might not be the culprit if his NVMe /dev nodes are moving around. RAID0 isn't the cause but it will make things more complicated when something fails further down in the stack. If his system is dynamically naming devices in /dev/nvme* then that needs to be dealt with before even thinking about RAID. Not really sure where to start looking at that off the top of my head since I was under the assumption that this wasn't supposed to happen with NVMe. There was recently a bit of discussion about LVM, and Rich sent me some links. I tried to read and understand it, but it seemed even more complicated than RAID. Plus, several years ago, when I was using Fedora, one of their obligatory updates changed my setup to LVM (without telling me that it was going to do so), and I couldn't get rid of it. That left a bad taste in my mouth and I have always avoided LVM ever since. But I must admit that my dislike of LVM is pure bias without much science. I am more concerned about devices renaming themselves and changing how they are mounted, all without any input from me. About January 20 I lost the first array that I had been running without a problem for about a month. And now my re-creation of that array is playing up after only a week. As I mentioned before, after rebooting the drives appear fine, read-write, but when I launched Ktorrent it complained that about half of the files it was seeding were missing. The files are all there and I can do anything I want to with them, but something is screwy with access. And why just half of the files? Either they should all work or they should all fail. That's what seems so odd. A defective drive wouldn't actually change the way things are enumerated. You have 4 drives, one of those would disappear and the others would stay the same (for the most part). A simple test to help everyone here understand what your machine is doing would be to run through a few reboots and grab the list of devices, like so 1) unplug your TB-3 drives and reboot. 2) record the output of 'ls -l /dev/nvme*' here 3) turn the computer off 4) plug in the TB-3 drives 5) turn the computer on and run 'ls /dev/nvme*' again. This will clearly isolate the device nodes for your enclosure independently of everything else on your computer. Once we have the drives isolate, it's trivial to watch them for irregular behavior. Until we have more confidence in the existence of your /dev/nvme nodes we can ignore the other symptoms. Right now my money is on a defective drive. I need to find some tools for diagnosis and learn how to use them. All four are brand new Intel drives. If they all check out OK, then it's time to consider other possibilities. ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Mon, 1 Feb 2021 23:48:03 -0800 Ben Koenig dijo: >On 2/1/21 11:35 PM, John Jason Jordan wrote: >> On Mon, 1 Feb 2021 22:15:12 -0800 >> Ben Koenig dijo: >> Perhaps now would be the time to dig out those old emails and consider some of the native alternatives rejected in favor of RAID0. >>> Unfortunately it looks like RAID might not be the culprit if his >>> NVMe /dev nodes are moving around. RAID0 isn't the cause but it >>> will make things more complicated when something fails further down >>> in the stack. >>> >>> If his system is dynamically naming devices in /dev/nvme* then that >>> needs to be dealt with before even thinking about RAID. Not really >>> sure where to start looking at that off the top of my head since I >>> was under the assumption that this wasn't supposed to happen with >>> NVMe. >> There was recently a bit of discussion about LVM, and Rich sent me >> some links. I tried to read and understand it, but it seemed even >> more complicated than RAID. Plus, several years ago, when I was >> using Fedora, one of their obligatory updates changed my setup to >> LVM (without telling me that it was going to do so), and I couldn't >> get rid of it. That left a bad taste in my mouth and I have always >> avoided LVM ever since. But I must admit that my dislike of LVM is >> pure bias without much science. >> >> I am more concerned about devices renaming themselves and changing >> how they are mounted, all without any input from me. About January >> 20 I lost the first array that I had been running without a problem >> for about a month. And now my re-creation of that array is playing >> up after only a week. As I mentioned before, after rebooting the >> drives appear fine, read-write, but when I launched Ktorrent it >> complained that about half of the files it was seeding were missing. >> The files are all there and I can do anything I want to with them, >> but something is screwy with access. And why just half of the files? >> Either they should all work or they should all fail. > >That's what seems so odd. A defective drive wouldn't actually change >the way things are enumerated. You have 4 drives, one of those would >disappear and the others would stay the same (for the most part). My understanding is that, since it is RAID0, if one drive fails the whole array fails. (But that's why this array is backed up to a NAS.) >A simple test to help everyone here understand what your machine is >doing would be to run through a few reboots and grab the list of >devices, like so I will do these things in the morning. It's too late and my brain is going into shutdown mode. But I should add one more thought: I have a RAID0 array on the Synology NAS, and another on a Mediasonic enclosure with two WS drives. Both have worked flawlessly for about four years. I've never had to mess with the arrays. And one more thought: I'd consider LVM instead of RAID0. But whatever system I set up, I need the four 7.68TB NVMe drives to appear as one big-ass 31TB drive. Now it's bedtime. :) >1) unplug your TB-3 drives and reboot. > >2) record the output of 'ls -l /dev/nvme*' here > >3) turn the computer off > >4) plug in the TB-3 drives > >5) turn the computer on and run 'ls /dev/nvme*' again. > > >This will clearly isolate the device nodes for your enclosure >independently of everything else on your computer. Once we have the >drives isolate, it's trivial to watch them for irregular behavior. >Until we have more confidence in the existence of your /dev/nvme nodes >we can ignore the other symptoms. ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Tue, 2 Feb 2021, John Jason Jordan wrote: And one more thought: I'd consider LVM instead of RAID0. But whatever system I set up, I need the four 7.68TB NVMe drives to appear as one big-ass 31TB drive. John, That's what LVM does. You have four physical drives. They can be collected into virtual groups any way you want, including all four in one group. Then each group (one in your case) can be divided into as many logical volumes as you want ... one in your case. That's the way I set up my Mediasonic 4-bay NAS. I now have a single 8TiB volume available. The process is a bit different from preparing a new drive (hard or solid state), but one of the URLs I sent lays them out step-by-step. So, should you choose to once again dip your toes in the LVM waters you'll find plenty of help here to keep them toes warm. Keep on truckin', Rich ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Mon, 1 Feb 2021 23:48:03 -0800 Ben Koenig dijo: >A simple test to help everyone here understand what your machine is >doing would be to run through a few reboots and grab the list of >devices, like so > >1) unplug your TB-3 drives and reboot. > >2) record the output of 'ls -l /dev/nvme*' here > >3) turn the computer off > >4) plug in the TB-3 drives > >5) turn the computer on and run 'ls /dev/nvme*' again. > >This will clearly isolate the device nodes for your enclosure >independently of everything else on your computer. Once we have the >drives isolate, it's trivial to watch them for irregular behavior. >Until we have more confidence in the existence of your /dev/nvme nodes >we can ignore the other symptoms. Here are the results: 1: (after unplugging TB3 device and rebooting) crw--- 1 root root 239, 0 Feb 2 12:01 /dev/nvme0 brw-rw 1 root disk 259, 0 Feb 2 12:01 /dev/nvme0n1 brw-rw 1 root disk 259, 1 Feb 2 12:01 /dev/nvme0n1p1 brw-rw 1 root disk 259, 2 Feb 2 12:01 /dev/nvme0n1p2 Note that nvme0 is a 1TB m.2 drive inside the Thinkpad that holds / and /home. 2: (after turning off computer, plugging in TB3 device, and booting) crw--- 1 root root 239, 0 Feb 2 11:47 /dev/nvme0 brw-rw 1 root disk 259, 0 Feb 2 11:47 /dev/nvme0n1 crw--- 1 root root 239, 1 Feb 2 11:47 /dev/nvme1 brw-rw 1 root disk 259, 2 Feb 2 11:47 /dev/nvme1n1 crw--- 1 root root 239, 2 Feb 2 11:47 /dev/nvme2 brw-rw 1 root disk 259, 1 Feb 2 11:47 /dev/nvme2n1 crw--- 1 root root 239, 3 Feb 2 11:47 /dev/nvme3 brw-rw 1 root disk 259, 3 Feb 2 11:47 /dev/nvme3n1 crw--- 1 root root 239, 4 Feb 2 11:47 /dev/nvme4 brw-rw 1 root disk 259, 4 Feb 2 11:47 /dev/nvme4n1 brw-rw 1 root disk 259, 5 Feb 2 11:47 /dev/nvme4n1p1 brw-rw 1 root disk 259, 6 Feb 2 11:47 /dev/nvme4n1p2 ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On 2/2/21 12:09 PM, John Jason Jordan wrote: On Mon, 1 Feb 2021 23:48:03 -0800 Ben Koenig dijo: A simple test to help everyone here understand what your machine is doing would be to run through a few reboots and grab the list of devices, like so 1) unplug your TB-3 drives and reboot. 2) record the output of 'ls -l /dev/nvme*' here 3) turn the computer off 4) plug in the TB-3 drives 5) turn the computer on and run 'ls /dev/nvme*' again. This will clearly isolate the device nodes for your enclosure independently of everything else on your computer. Once we have the drives isolate, it's trivial to watch them for irregular behavior. Until we have more confidence in the existence of your /dev/nvme nodes we can ignore the other symptoms. Here are the results: 1: (after unplugging TB3 device and rebooting) crw--- 1 root root 239, 0 Feb 2 12:01 /dev/nvme0 brw-rw 1 root disk 259, 0 Feb 2 12:01 /dev/nvme0n1 brw-rw 1 root disk 259, 1 Feb 2 12:01 /dev/nvme0n1p1 brw-rw 1 root disk 259, 2 Feb 2 12:01 /dev/nvme0n1p2 Note that nvme0 is a 1TB m.2 drive inside the Thinkpad that holds / and /home. 2: (after turning off computer, plugging in TB3 device, and booting) crw--- 1 root root 239, 0 Feb 2 11:47 /dev/nvme0 brw-rw 1 root disk 259, 0 Feb 2 11:47 /dev/nvme0n1 crw--- 1 root root 239, 1 Feb 2 11:47 /dev/nvme1 brw-rw 1 root disk 259, 2 Feb 2 11:47 /dev/nvme1n1 crw--- 1 root root 239, 2 Feb 2 11:47 /dev/nvme2 brw-rw 1 root disk 259, 1 Feb 2 11:47 /dev/nvme2n1 crw--- 1 root root 239, 3 Feb 2 11:47 /dev/nvme3 brw-rw 1 root disk 259, 3 Feb 2 11:47 /dev/nvme3n1 crw--- 1 root root 239, 4 Feb 2 11:47 /dev/nvme4 brw-rw 1 root disk 259, 4 Feb 2 11:47 /dev/nvme4n1 brw-rw 1 root disk 259, 5 Feb 2 11:47 /dev/nvme4n1p1 brw-rw 1 root disk 259, 6 Feb 2 11:47 /dev/nvme4n1p2 OK so then everything seems to be connecting at the hardware level. Your TB controller is exposing 4 NVMe devices and they are identifying as block devices ("disks") which means that the hardware is functioning. What stands out is that of the 4 disks only 1 of them actually has partitions. This strikes me as odd. As a simple test, can you create a test folder in /mnt and see if those partitions on nvme4 are mountable? $ mkdir /mnt/nvme $ mount /dev/nvme4n1p1 /mnt/nvme Both commands should be run as root and I'm assuming you have nothing mounted in /mnt. If that succeeds and you can view files let me know. Also post any output from the mount command here. -Ben ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
I have used F2FS for years on my SSD drives. [F2FS ALWAYS benchmarks FASTEST fs] There is an option in mkfs.f2fs -c , that can make use of upto 7 drives as ONE LARGE DRIVE. Have 7 240G [do not have to be same size] as one /dev in one of my large boxes i use archlinux, YMMV On 2/2/21, Rich Shepard wrote: > On Tue, 2 Feb 2021, John Jason Jordan wrote: > >> And one more thought: I'd consider LVM instead of RAID0. But whatever >> system I set up, I need the four 7.68TB NVMe drives to appear as one >> big-ass 31TB drive. > > John, > > That's what LVM does. You have four physical drives. They can be collected > into virtual groups any way you want, including all four in one group. Then > each group (one in your case) can be divided into as many logical volumes > as > you want ... one in your case. > > That's the way I set up my Mediasonic 4-bay NAS. I now have a single 8TiB > volume available. > > The process is a bit different from preparing a new drive (hard or solid > state), but one of the URLs I sent lays them out step-by-step. > > So, should you choose to once again dip your toes in the LVM waters you'll > find plenty of help here to keep them toes warm. > > Keep on truckin', > > Rich > ___ > PLUG: https://pdxlinux.org > PLUG mailing list > PLUG@pdxlinux.org > http://lists.pdxlinux.org/mailman/listinfo/plug > ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Tue, 2 Feb 2021 12:09:06 -0800 John Jason Jordan dijo: >On Mon, 1 Feb 2021 23:48:03 -0800 >Ben Koenig dijo: > >>A simple test to help everyone here understand what your machine is >>doing would be to run through a few reboots and grab the list of >>devices, like so >> >>1) unplug your TB-3 drives and reboot. >> >>2) record the output of 'ls -l /dev/nvme*' here >> >>3) turn the computer off >> >>4) plug in the TB-3 drives >> >>5) turn the computer on and run 'ls /dev/nvme*' again. >> >>This will clearly isolate the device nodes for your enclosure >>independently of everything else on your computer. Once we have the >>drives isolate, it's trivial to watch them for irregular behavior. >>Until we have more confidence in the existence of your /dev/nvme nodes >>we can ignore the other symptoms. > >Here are the results: > >1: (after unplugging TB3 device and rebooting) >crw--- 1 root root 239, 0 Feb 2 12:01 /dev/nvme0 >brw-rw 1 root disk 259, 0 Feb 2 12:01 /dev/nvme0n1 >brw-rw 1 root disk 259, 1 Feb 2 12:01 /dev/nvme0n1p1 >brw-rw 1 root disk 259, 2 Feb 2 12:01 /dev/nvme0n1p2 >Note that nvme0 is a 1TB m.2 drive inside the Thinkpad that holds / and >/home. > >2: (after turning off computer, plugging in TB3 device, and booting) >crw--- 1 root root 239, 0 Feb 2 11:47 /dev/nvme0 >brw-rw 1 root disk 259, 0 Feb 2 11:47 /dev/nvme0n1 >crw--- 1 root root 239, 1 Feb 2 11:47 >/dev/nvme1 brw-rw 1 root disk 259, 2 Feb 2 11:47 /dev/nvme1n1 >crw--- 1 root root 239, 2 Feb 2 11:47 /dev/nvme2 >brw-rw 1 root disk 259, 1 Feb 2 11:47 /dev/nvme2n1 >crw--- 1 root root 239, 3 Feb 2 11:47 /dev/nvme3 >brw-rw 1 root disk 259, 3 Feb 2 11:47 /dev/nvme3n1 >crw--- 1 root root 239, 4 Feb 2 11:47 /dev/nvme4 >brw-rw 1 root disk 259, 4 Feb 2 11:47 /dev/nvme4n1 >brw-rw 1 root disk 259, 5 Feb 2 11:47 /dev/nvme4n1p1 >brw-rw 1 root disk 259, 6 Feb 2 11:47 /dev/nvme4n1p2 After the above I opened Ktorrent. It presented me with a couple error messages about missing files, where I pointed it to a folder that I had renamed, which it happily accepted, and it is now seeding all its torrents. The renamed folders were my fault. Then I opened a file manager to /dev and scrolled down to nvme entries; which gave me: nvme0 nvme0n1 nvme1 nvme1n1 nvme2 nvme2n1 nvme3 nvme3n1 nvme4 nvme4n1 nvme4n1p1 nvme4n1p2 And scrolling up a bit I see md127 and md127p1. Everything is back to normal. My only problem is what happens when the md127 and md127p1 suddenly become read-only again. It happened during the night of February 1, so I can assume that eventually it's going to happen again. ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On 2/2/21 12:31 PM, John Jason Jordan wrote: On Tue, 2 Feb 2021 12:09:06 -0800 John Jason Jordan dijo: On Mon, 1 Feb 2021 23:48:03 -0800 Ben Koenig dijo: A simple test to help everyone here understand what your machine is doing would be to run through a few reboots and grab the list of devices, like so 1) unplug your TB-3 drives and reboot. 2) record the output of 'ls -l /dev/nvme*' here 3) turn the computer off 4) plug in the TB-3 drives 5) turn the computer on and run 'ls /dev/nvme*' again. This will clearly isolate the device nodes for your enclosure independently of everything else on your computer. Once we have the drives isolate, it's trivial to watch them for irregular behavior. Until we have more confidence in the existence of your /dev/nvme nodes we can ignore the other symptoms. Here are the results: 1: (after unplugging TB3 device and rebooting) crw--- 1 root root 239, 0 Feb 2 12:01 /dev/nvme0 brw-rw 1 root disk 259, 0 Feb 2 12:01 /dev/nvme0n1 brw-rw 1 root disk 259, 1 Feb 2 12:01 /dev/nvme0n1p1 brw-rw 1 root disk 259, 2 Feb 2 12:01 /dev/nvme0n1p2 Note that nvme0 is a 1TB m.2 drive inside the Thinkpad that holds / and /home. 2: (after turning off computer, plugging in TB3 device, and booting) crw--- 1 root root 239, 0 Feb 2 11:47 /dev/nvme0 brw-rw 1 root disk 259, 0 Feb 2 11:47 /dev/nvme0n1 crw--- 1 root root 239, 1 Feb 2 11:47 /dev/nvme1 brw-rw 1 root disk 259, 2 Feb 2 11:47 /dev/nvme1n1 crw--- 1 root root 239, 2 Feb 2 11:47 /dev/nvme2 brw-rw 1 root disk 259, 1 Feb 2 11:47 /dev/nvme2n1 crw--- 1 root root 239, 3 Feb 2 11:47 /dev/nvme3 brw-rw 1 root disk 259, 3 Feb 2 11:47 /dev/nvme3n1 crw--- 1 root root 239, 4 Feb 2 11:47 /dev/nvme4 brw-rw 1 root disk 259, 4 Feb 2 11:47 /dev/nvme4n1 brw-rw 1 root disk 259, 5 Feb 2 11:47 /dev/nvme4n1p1 brw-rw 1 root disk 259, 6 Feb 2 11:47 /dev/nvme4n1p2 After the above I opened Ktorrent. It presented me with a couple error messages about missing files, where I pointed it to a folder that I had renamed, which it happily accepted, and it is now seeding all its torrents. The renamed folders were my fault. Then I opened a file manager to /dev and scrolled down to nvme entries; which gave me: nvme0 nvme0n1 nvme1 nvme1n1 nvme2 nvme2n1 nvme3 nvme3n1 nvme4 nvme4n1 nvme4n1p1 nvme4n1p2 And scrolling up a bit I see md127 and md127p1. Everything is back to normal. My only problem is what happens when the md127 and md127p1 suddenly become read-only again. It happened during the night of February 1, so I can assume that eventually it's going to happen again. There's probably a bunch of debug information dumped into a log somewhere. If it seems to be working now and you didn't actually change anything then it was probably a minor drop in connection with one or more of the drives. Depending on how paranoid you want to be there are steps you can take to try and root cause the problem. The one thing to keep an eye on is the number of nvme devices in /dev. If you encounter a situation where you have drives numbered above nvme4 then that means the kernel is losing track of your connected drives. That's a very specific class of problem and usually easy to deal with once isolated. For now it's probably best to continue as normal and if it happens again just take a look at your nvme devices in /dev and see if they have changed. -Ben ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Tue, 2 Feb 2021 05:50:57 -0800 (PST) Rich Shepard dijo: >On Tue, 2 Feb 2021, John Jason Jordan wrote: > >> And one more thought: I'd consider LVM instead of RAID0. But whatever >> system I set up, I need the four 7.68TB NVMe drives to appear as one >> big-ass 31TB drive. >That's what LVM does. You have four physical drives. They can be >collected into virtual groups any way you want, including all four in >one group. Then each group (one in your case) can be divided into as >many logical volumes as you want ... one in your case. > >That's the way I set up my Mediasonic 4-bay NAS. I now have a single >8TiB volume available. > >The process is a bit different from preparing a new drive (hard or >solid state), but one of the URLs I sent lays them out step-by-step. At the moment everything is working. But I'm pretty sure things will eventually turn upside down again. My best guess is that one of my drives is defective, but the failure is intermittent. I need something to check the disks, sort of like the memtest tool that appears in the grub menu. I want something that will check every part of each drive, repeatedly. I haven't completely given up on LVM, and I may end up there eventually. But while poking around I stumbled on F2FS, a Linux filesystem for flash drives written by Samsung. Here is the Wikipedia article: https://en.wikipedia.org/wiki/F2FS Opinions welcome! ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On 2/2/21 3:20 PM, John Jason Jordan wrote: On Tue, 2 Feb 2021 05:50:57 -0800 (PST) Rich Shepard dijo: On Tue, 2 Feb 2021, John Jason Jordan wrote: And one more thought: I'd consider LVM instead of RAID0. But whatever system I set up, I need the four 7.68TB NVMe drives to appear as one big-ass 31TB drive. That's what LVM does. You have four physical drives. They can be collected into virtual groups any way you want, including all four in one group. Then each group (one in your case) can be divided into as many logical volumes as you want ... one in your case. That's the way I set up my Mediasonic 4-bay NAS. I now have a single 8TiB volume available. The process is a bit different from preparing a new drive (hard or solid state), but one of the URLs I sent lays them out step-by-step. At the moment everything is working. But I'm pretty sure things will eventually turn upside down again. My best guess is that one of my drives is defective, but the failure is intermittent. I need something to check the disks, sort of like the memtest tool that appears in the grub menu. I want something that will check every part of each drive, repeatedly. smartctl can give you a quick look at drive statistics. $ smartctl -a /dev/nvme0n1 It's by no means guaranteed to tell you if something is wrong or what the cause is, but it will probably let you know if there is a major problem. I haven't completely given up on LVM, and I may end up there eventually. But while poking around I stumbled on F2FS, a Linux filesystem for flash drives written by Samsung. Here is the Wikipedia article: https://en.wikipedia.org/wiki/F2FS Opinions welcome! ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Tue, 2 Feb 2021 17:07:45 -0800 Ben Koenig dijo: > smartctl -a /dev/nvme0n1 OK, this is really weird. All this time nvme0 has been the 1TB m.2 Samsung drive inside the Thinkpad, which holds / and /home. I was amazed when I ran the command exactly as above and it said the drive was a 7.68TB Intel NVMe. I continued, incrementing the number each time I ran the command, and the 1TB Samsung is now nvme4n1. I had been perplexed with why nvme4n1 had two partitions, because I was thinking it was part of the array. The mdadm command that I used to create the array listed nvme1, nvme2, nvme3, and nvme4, in that order. I copied the command from the terminal and put it in a text file to save it, and I verified just now that I did not misremember. Drive nvme4 was originally part of the array, and now it is the m.2 drive, while the array is now made of nvme0-3. I have long been aware that drive labeling like sda, sdb, etc. can annoyingly swap around. Apparently NVMe drives can be just as exasperating. As for the results of the command on the drives in the array, there were no errors reported, and everything else looked normal. ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Tue, Feb 2, 2021 at 6:32 PM John Jason Jordan wrote: > I have long been aware that drive labeling like sda, sdb, etc. can > annoyingly swap around. Apparently NVMe drives can be just as > exasperating. > > Do you have a /dev/disk directory? If so, within it you should find various ways of addressing each drive, including by UUID. You could specify this path in your fstab file, as well as re-create the RAID array using these names rather than /dev/nvmewhatever. -wes ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Tue, 2 Feb 2021 18:49:21 -0800 wes dijo: >On Tue, Feb 2, 2021 at 6:32 PM John Jason Jordan >wrote: > >> I have long been aware that drive labeling like sda, sdb, etc. can >> annoyingly swap around. Apparently NVMe drives can be just as >> exasperating. >Do you have a /dev/disk directory? If so, within it you should find >various ways of addressing each drive, including by UUID. You could >specify this path in your fstab file, as well as re-create the RAID >array using these names rather than /dev/nvmewhatever. Oh, I always do. After creating the array as md0 I opened it in Gparted (GUI) to create and format the partition. If you right-click on the partition you can get its UUID, which I saved in my text file where I keep all the junk I had to go through to create the array and mount it. It's nice to know that you can get the UUID from /dev/disk, but there is also a command (that I have forgotten) to get the UUID. For a long time I used LABEL= for mounting, where I had give the partition a label. Labels are great, but you have to be careful not to label more than one partition with the same label. I like labels, because they are human readable, unlike UUIDs. ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Tue, Feb 2, 2021 at 7:36 PM John Jason Jordan wrote: > It's nice to know that you can get the UUID from /dev/disk, but there > is also a command (that I have forgotten) to get the UUID. For a long > time I used LABEL= for mounting, where I had give the partition a > label. Labels are great, but you have to be careful not to label more > than one partition with the same label. I like labels, because they are > human readable, unlike UUIDs. > > If you create the RAID array with UUIDs, it (probably) won't break when your system renames devices. -wes ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Tue, 2 Feb 2021 19:46:51 -0800 wes dijo: >If you create the RAID array with UUIDs, it (probably) won't break when >your system renames devices. Brilliant! ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
You can use mdadm to examine the superblock on each drive and it will give you the details of the array that the drive thinks it is in. Without a mdadm.conf, the kernel will attempt to assemble the array based on what it finds in the drive superblocks and will default to md127 and count up from there. /proc/mdadm will give you the status of the assembled arrays directly without the need to go through the mdadm util. On Tue, Feb 2, 2021 at 9:09 PM John Jason Jordan wrote: > > On Tue, 2 Feb 2021 19:46:51 -0800 > wes dijo: > > >If you create the RAID array with UUIDs, it (probably) won't break when > >your system renames devices. > > Brilliant! > ___ > PLUG: https://pdxlinux.org > PLUG mailing list > PLUG@pdxlinux.org > http://lists.pdxlinux.org/mailman/listinfo/plug ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Wed, 3 Feb 2021 12:48:48 -0800 Larry Brigman dijo: >You can use mdadm to examine the superblock on each drive and it will >give you the details of the array that the drive thinks it is in. >Without a mdadm.conf, the kernel will attempt to assemble the array >based on what it finds in the drive superblocks and will default to >md127 and count up from there. Ah, now I know where the '127' came from. As for mdadm.conf, when I started I noticed that I had such a file, and I left it alone. But when I got massive errors in my attempts to create the array, in hopes of making things work better, I appended '.old' to the file. I still had errors, but after trying over and over I finally succeeded in creating /dev/md0 without errors. After I finished I deleted the renamed mdadm.conf file, and then used: sudo mdadm --detail --scan | sudo tee -a /etc/mdadm/mdadm.conf to recreate the mdadm.conf file. >/proc/mdadm will give you the status of the assembled arrays directly >without the need to go through the mdadm util. /proc/mdadm bash: /proc/mdadm: No such file or directory ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Wed, Feb 3, 2021 at 2:19 PM John Jason Jordan wrote: > On Wed, 3 Feb 2021 12:48:48 -0800 Larry Brigman > dijo: > >/proc/mdadm will give you the status of the assembled arrays directly > >without the need to go through the mdadm util. > > /proc/mdadm > bash: /proc/mdadm: No such file or directory > > I get this mixed up frequently also. it's actually /proc/mdstat. -wes ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Wed, 3 Feb 2021 15:05:33 -0800 wes dijo: >On Wed, Feb 3, 2021 at 2:19 PM John Jason Jordan >wrote: > >> On Wed, 3 Feb 2021 12:48:48 -0800 Larry Brigman >> dijo: >> >/proc/mdadm will give you the status of the assembled arrays >> >directly without the need to go through the mdadm util. >> >> /proc/mdadm >> bash: /proc/mdadm: No such file or directory >I get this mixed up frequently also. it's actually /proc/mdstat. $ /proc/mdadm bash: /proc/mdadm: No such file or directory jjj@Devil-Thinkpad:~$ /proc/mdstat bash: /proc/mdstat: Permission denied jjj@Devil-Thinkpad:~$ sudo /proc/mdstat [sudo] password for jjj: sudo: /proc/mdstat: command not found Some utility for mdadm that I don't have installed? ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Wed, Feb 3, 2021 at 4:14 PM John Jason Jordan wrote: > > $ /proc/mdstat > > Ah, yeah, we could have been clearer about that too. try cat /proc/mdstat -wes ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Wed, 3 Feb 2021, John Jason Jordan wrote: jjj@Devil-Thinkpad:~$ sudo /proc/mdstat John, The /proc/files are text. Use: sudo less /proc/mdstat Rich ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Tue, 2 Feb 2021 12:31:45 -0800 John Jason Jordan dijo: >nvme0 >nvme0n1 >nvme1 >nvme1n1 >nvme2 >nvme2n1 >nvme3 >nvme3n1 >nvme4 >nvme4n1 >nvme4n1p1 >nvme4n1p2 > >And scrolling up a bit I see md127 and md127p1. > >Everything is back to normal. My only problem is what happens when the >md127 and md127p1 suddenly become read-only again. It happened during >the night of February 1, so I can assume that eventually it's going to >happen again. It worked fine for a couple weeks, and now it has stopped working again. The contents of /dev now include: md127 md127p1 nvme0n2 #was nvme0n1 nvme1n2 #was nvme1n1 nvme2n2 #was nvme2n1 nvme3n2 #was nvme3n1 nvme4 nvme4n1 nvme4n1p1 nvme4n1p2 nvme5 nvme6 nvme7 nvme8 And nvme5-8 do not exist. Things are all screwed up again. I am unable to access any of the files on md127p1 (the array). The array is (supposed to be) made up of nvme0-3, and nvme4 is inside my Thinkpad for / and /home. I pulled the TB3 cable out of the enclosure (which automatically shuts it down) and then plugged it back in again. Nothing changed. From past experience the only way to restore things is to completely reboot. I really need to figure out what is causing this. Is it a defective drive? Is it one of the two PCI cards that the four drives are plugged into? Is it the enclosure that the two PCI cards are plugged into? Why does it work for a couple weeks and then go south? The only thing that I can rule out is the TB3 cable, because I bought three of them and they have been swapped, with no change. This is driving me nuts. ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Wed, 17 Feb 2021 14:05:20 -0800 John Jason Jordan dijo: >This is driving me nuts. Before rebooting I tried to umount md127p1, but got the 'busy' error message, which I solved with the 'lazy' -l option. But although umount -l executed without error, it didn't umount the array. Eventually I gave up and rebooted, and after rebooting everything is back to normal. I really need ideas. ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Wed, Feb 17, 2021, 17:30 John Jason Jordan wrote: > On Wed, 17 Feb 2021 14:05:20 -0800 > John Jason Jordan dijo: > > >This is driving me nuts. > > Before rebooting I tried to umount md127p1, but got the 'busy' error > message, which I solved with the 'lazy' -l option. But although umount > -l executed without error, it didn't umount the array. Eventually I > gave up and rebooted, and after rebooting everything is back to normal. > > I really need ideas. > I thought you switched to using uuids - why do you care about name changes? TB is removable hotplug like USB ... device name changes are to be expected. No? ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Wed, 17 Feb 2021 18:55:11 -0500 Tomas Kuchta dijo: >On Wed, Feb 17, 2021, 17:30 John Jason Jordan wrote: > >> On Wed, 17 Feb 2021 14:05:20 -0800 >> John Jason Jordan dijo: >> >> >This is driving me nuts. >> >> Before rebooting I tried to umount md127p1, but got the 'busy' error >> message, which I solved with the 'lazy' -l option. But although >> umount -l executed without error, it didn't umount the array. >> Eventually I gave up and rebooted, and after rebooting everything is >> back to normal. >> >> I really need ideas. >I thought you switched to using uuids - why do you care about name >changes? > >TB is removable hotplug like USB ... device name changes are to be >expected. Well, yes. But I spent just a couple minute trying to umount and mount the array, and it didn't dawn on my to use the UUID. I had a file browser window open and /dev was displayed, so I just used /dev/md127p1. That should have worked, at least for umount. When I rebooted the array was mounted with the UUID, because the line in fstab has the UUID, not /dev/md127. The following command executed without error, but afterwards the array was still mounted, even though I was unable to access any of the files on it: sudo umount -l /dev/md127p1 And afterwards I tried to mount it, but the mount command gave the error that the device was already mounted. And before umount/mount I pulled the TB3 plug from the enclosure, waited a couple of minutes, then plugged it back in, but I was still unable to access any of the files. It should have automatically mounted, because that's what TB3 does. From past history I was pretty sure that rebooting would fix things - I was just trying to solve the problem without having to reboot. And none of these weird results surprised me, because when I discovered that I couldn't access files on the array my very first step was to look at /dev, and when I saw the strange list of device names I knew that things had gotten screwed up again. Yes, device names may change on booting or when plugged in, but don't device names stay put once you are booted or a drive is plugged in? ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Wed, Feb 17, 2021, 19:41 John Jason Jordan wrote: > On Wed, 17 Feb 2021 18:55:11 -0500 > Tomas Kuchta dijo: > > >On Wed, Feb 17, 2021, 17:30 John Jason Jordan wrote: > > > >> On Wed, 17 Feb 2021 14:05:20 -0800 > >> John Jason Jordan dijo: > >> > >> >This is driving me nuts. > >> > >> Before rebooting I tried to umount md127p1, but got the 'busy' error > >> message, which I solved with the 'lazy' -l option. But although > >> umount -l executed without error, it didn't umount the array. > >> Eventually I gave up and rebooted, and after rebooting everything is > >> back to normal. > >> > >> I really need ideas. > > >I thought you switched to using uuids - why do you care about name > >changes? > > > >TB is removable hotplug like USB ... device name changes are to be > >expected. > > Well, yes. But I spent just a couple minute trying to umount and mount > the array, and it didn't dawn on my to use the UUID. I had a file > browser window open and /dev was displayed, so I just used > /dev/md127p1. That should have worked, at least for umount. When I > rebooted the array was mounted with the UUID, because the line in fstab > has the UUID, not /dev/md127. The following command executed without > error, but afterwards the array was still mounted, even though I was > unable to access any of the files on it: > > sudo umount -l /dev/md127p1 > > And afterwards I tried to mount it, but the mount command gave the error > that the device was already mounted. > > And before umount/mount I pulled the TB3 plug from the enclosure, > waited a couple of minutes, then plugged it back in, but I was still > unable to access any of the files. It should have automatically > mounted, because that's what TB3 does. That might work with single disk if you run sync beforehand and nothing accesses the disk. Definitely not a good idea to unplug disk array without clean unmount and stopping mdadm. You say that in the past was great. Your past and current problems would indicate that it was not all great in the past - or you runout of luck. Hope that helps, T > > ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Thu, 18 Feb 2021 00:27:25 -0500 Tomas Kuchta dijo: >> And before umount/mount I pulled the TB3 plug from the enclosure, >> waited a couple of minutes, then plugged it back in, but I was still >> unable to access any of the files. It should have automatically >> mounted, because that's what TB3 does. >That might work with single disk if you run sync beforehand and nothing >accesses the disk. I never heard of 'sync.' Is that a command? >Definitely not a good idea to unplug disk array without clean unmount >and stopping mdadm. I have had numerous RAID0 arrays, and I never had to stop mdadm before mounting/unmounting. I don't even know how to stop mdadm. >You say that in the past was great. Your past and current problems >would indicate that it was not all great in the past - or you runout of >luck. I think I need some education. An mdadm array is apparently not the same as a disk, but how it's different and how to deal with it is something I need to learn about. ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Wed, 2021-02-17 at 21:59 -0800, John Jason Jordan wrote: > On Thu, 18 Feb 2021 00:27:25 -0500 > Tomas Kuchta dijo: > > > > And before umount/mount I pulled the TB3 plug from the enclosure, > > > waited a couple of minutes, then plugged it back in, but I was > > > still > > > unable to access any of the files. It should have automatically > > > mounted, because that's what TB3 does. > > That might work with single disk if you run sync beforehand and > > nothing > > accesses the disk. > > I never heard of 'sync.' Is that a command? > > > Definitely not a good idea to unplug disk array without clean > > unmount > > and stopping mdadm. > > I have had numerous RAID0 arrays, and I never had to stop mdadm > before > mounting/unmounting. I don't even know how to stop mdadm. > > > You say that in the past was great. Your past and current > > problems > > would indicate that it was not all great in the past - or you > > runout of > > luck. > > I think I need some education. An mdadm array is apparently not the > same > as a disk, but how it's different and how to deal with it is > something > I need to learn about. > Note: My advice is: Never let the array or TB go to sleep - SSDs can sleep to save power if mdraid lets them. There are two ways to cleanly stop raid array: 1. Easiest: sudo shutdown -h 2. If you need to change configuration, add/remove disks, etc. Copy paste from: -Tomas ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Wed, 17 Feb 2021 22:57:55 -0800 TomasK dijo: >On Wed, 2021-02-17 at 21:59 -0800, John Jason Jordan wrote: >> On Thu, 18 Feb 2021 00:27:25 -0500 >> Tomas Kuchta dijo: >> I never heard of 'sync.' Is that a command? >> >> > Definitely not a good idea to unplug disk array without clean >> > unmount >> > and stopping mdadm. >Note: My advice is: Never let the array or TB go to sleep - SSDs can >sleep to save power if mdraid lets them. I need to figure out how to stop SSDs from sleep to save power. >There are two ways to cleanly stop raid array: > 1. Easiest: > sudo shutdown -h > 2. If you need to change configuration, add/remove disks, etc. > Copy > paste from: What does -h reference? How do I enrer a command to showdown the arrary> ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Thu, Feb 18, 2021, 02:18 John Jason Jordan wrote: > > > >There are two ways to cleanly stop raid array: > > 1. Easiest: > > sudo shutdown -h > > 2. If you need to change configuration, add/remove disks, etc. > > Copy > > paste from: > > What does -h reference? How do I enrer a command to showdown the > arrary> > . sudo shutdown -h now Will shutdown your laptop - there are, obviously, other ways of doing that - the shutdown process should take care of doing it correctly - as long as the numbers/dependencies in your fstab are correct. man shutdown Should answer all the questions about shutdown command options. -Tomas > ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On 2/17/21 4:40 PM, John Jason Jordan wrote: Yes, device names may change on booting or when plugged in, but don't device names stay put once you are booted or a drive is plugged in? This sounds like a cable or hardware issue to me. If a drive is mounted and something happens and the kernel isn't involved, you will get a new mount address of the same thing. Either the drives are going to sleep, as already suggested, or there is a glitch in the communication path which is causing a sporadic disconnect/reconnect sequence. I have seen this periodically with USB, and TB3 likely uses something similar, where udev is involved. Sorry that I don't have a solution, but if you have another TB3 cable you could use, that may help narrow things down if the problem goes away. dafr ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Thu, 18 Feb 2021 10:40:04 -0800 David dijo: >On 2/17/21 4:40 PM, John Jason Jordan wrote: > >> Yes, device names may change on booting or when plugged in, but don't >> device names stay put once you are booted or a drive is plugged in? > >This sounds like a cable or hardware issue to me. If a drive is >mounted and something happens and the kernel isn't involved, you will >get a new mount address of the same thing. > >Either the drives are going to sleep, as already suggested, or there >is a glitch in the communication path which is causing a sporadic >disconnect/reconnect sequence. > >I have seen this periodically with USB, and TB3 likely uses something >similar, where udev is involved. > >Sorry that I don't have a solution, but if you have another TB3 cable >you could use, that may help narrow things down if the problem goes >away. It's definitely not the cable. I have three TB3 cables, and they have been swapped, yet the problem remains. The four U.2 drives are plugged into two PCI cards (two drives per card) which are then plugged into a four-slot enclosure. The enclosure has two TB3 ports, but no on-off switch. When a powered TB3 cable is plugged into either port the enclosure automatically powers on, and it powers off when the cable is unplugged. And its fan is loud enough that there is no mistaking when it is running. It is possible that there was a power loss of only a few milliseconds - enough to un-mount the array, but not long enough that the fan stopped running. But since it is TB3, when the power is restored, why doesn't the array automatically mount? Besides, looking at the NVMe drives in /dev, when things go upside down I see nvme5 to nvme8 listed, and those drives don't exist. After rebooting they are no longer listed. The array is made from nvme0 to nvme3, and nvme4 is inside my laptop and holds partitions for / and /home. And from the command line I get screwy results from the mount command - just 'mount' does not list the array (i.e., it must not be mounted), but if I try to mount it I get an error message that it is already mounted. Yet the file manager does not see it, nor does the ls command. ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Thu, Feb 18, 2021 at 1:24 PM John Jason Jordan wrote: > > It's definitely not the cable. I have three TB3 cables, and they have > been swapped, yet the problem remains. > > So the next logical step is to swap the other bits of hardware. Which is really difficult when you only have 1 of each. Normally in such a case we could reach out to the community to see if anyone has similar hardware we can borrow for troubleshooting, but I think you're the only one in the Portland metro area with one of those enclosures. > The four U.2 drives are plugged into two PCI cards (two drives per > card) which are then plugged into a four-slot enclosure. The enclosure > has two TB3 ports, but no on-off switch. When a powered TB3 cable is > plugged into either port the enclosure automatically powers on, and it > powers off when the cable is unplugged. > Does the enclosure offer any options to supply power externally? Might be worth a try, just as an experiment. -wes ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug
Re: [PLUG] Raving mad RAID
On Thu, 18 Feb 2021 13:36:15 -0800 wes dijo: >On Thu, Feb 18, 2021 at 1:24 PM John Jason Jordan >wrote: > >> >> It's definitely not the cable. I have three TB3 cables, and they have >> been swapped, yet the problem remains. >So the next logical step is to swap the other bits of hardware. Which >is really difficult when you only have 1 of each. Normally in such a >case we could reach out to the community to see if anyone has similar >hardware we can borrow for troubleshooting, but I think you're the >only one in the Portland metro area with one of those enclosures. >> The four U.2 drives are plugged into two PCI cards (two drives per >> card) which are then plugged into a four-slot enclosure. The >> enclosure has two TB3 ports, but no on-off switch. When a powered >> TB3 cable is plugged into either port the enclosure automatically >> powers on, and it powers off when the cable is unplugged. >Does the enclosure offer any options to supply power externally? Might >be worth a try, just as an experiment. It's a Magma Expressbox 3T-V3: https://www.onestopsystems.com/product/expressbox-3t-v3 Magma was bought out by One Stop Systems. In the pic on the above link it looks like there is a power button on the front, but it's really just a light. Inside there are four PCI slots, but one is occupied by the TB3 card, and one of the others is blocked by a big heat sink so you can't hang a long card off the end of it. But that's OK because I just need slots for two full length cards. There are also two fans, one in front blowing in and another in back blowing out. It has an external power supply that plugs in with a connector that looks like the end of a machine power cord, like the end of the power cord for a desktop PC, except that the end that plugs into the wall has a brick on it. I don't know how many volts the enclosure takes, but it's not straight 120v, because I assume the brick is stepping it down. So there is another way o turn it off - just pull the power cord out of its socket. But unlike the TB3 cord, plugging the power cord back in doesn't turn it on. Even if I had spares to swap for the other bits, it would be difficult to swap things out because it runs fine for a couple weeks before it fails. ___ PLUG: https://pdxlinux.org PLUG mailing list PLUG@pdxlinux.org http://lists.pdxlinux.org/mailman/listinfo/plug