Veah

Hi all, My server stopped responding, VMs unreachable, shares inaccessible.... basically dead. Hard shutdown, restarted, parity going and all is normal. Great. This is an opportunity for me to get better at reading and understanding the logs while I am trying to piece together the picture. Log section at bottom. What I believe happened was I had opened the case at some point in time which may have been when it crashed, but I do not know for sure. Cables may have wiggled while I was removing an unused sff8643 pigtail. Hoping for log reading tips. Enjoy below. Sep 19 08:47:25 Svr emhttpd: spinning down /dev/sdf Sep 19 09:08:57 Svr emhttpd: spinning down /dev/sdd Sep 19 09:55:03 Svr emhttpd: spinning down /dev/sdb Sep 19 10:08:39 Svr emhttpd: read SMART /dev/sdb Sep 19 10:08:53 Svr emhttpd: read SMART /dev/sdg Sep 19 10:09:08 Svr emhttpd: read SMART /dev/sde Sep 19 10:10:03 Svr emhttpd: read SMART /dev/sdc Sep 19 10:16:05 Svr emhttpd: read SMART /dev/sdd Sep 19 10:16:57 Svr kernel: ahci 0000:48:00.0: AHCI controller unavailable! Sep 19 10:16:57 Svr kernel: ata10: failed to resume link (SControl FFFFFFFF) Sep 19 10:16:57 Svr kernel: ata10: SATA link down (SStatus FFFFFFFF SControl FFFFFFFF) Sep 19 10:16:57 Svr kernel: ahci 0000:48:00.0: AHCI controller unavailable! Sep 19 10:16:57 Svr kernel: ata10: failed to resume link (SControl FFFFFFFF) Sep 19 10:16:57 Svr kernel: ata10: SATA link down (SStatus FFFFFFFF SControl FFFFFFFF) Sep 19 10:16:57 Svr kernel: ata10: limiting SATA link speed to <unknown> Sep 19 10:16:57 Svr kernel: ahci 0000:48:00.0: AHCI controller unavailable! Sep 19 10:16:57 Svr kernel: xhci_hcd 0000:46:00.1: xHCI host controller not responding, assume dead Sep 19 10:16:57 Svr kernel: xhci_hcd 0000:46:00.1: HC died; cleaning up Sep 19 10:16:57 Svr kernel: hid-generic 0003:051D:0003.0001: usb_submit_urb(ctrl) failed: -19 Sep 19 10:16:57 Svr kernel: usb 5-4: USB disconnect, device number 2 Sep 19 10:16:57 Svr kernel: usb 5-4.1: USB disconnect, device number 4 Sep 19 10:16:57 Svr kernel: usb 5-4.2: USB disconnect, device number 6 Sep 19 10:16:57 Svr kernel: device offline error, dev sda, sector 3742504 op 0x0:(READ) flags 0x80700 phys_seg 4 prio class 2 Sep 19 10:16:57 Svr kernel: device offline error, dev sda, sector 3742504 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2 Sep 19 10:16:57 Svr kernel: I/O error, dev loop1, sector 54336 op 0x0:(READ) flags 0x800 phys_seg 13 prio class 2 Sep 19 10:16:57 Svr kernel: SQUASHFS error: Failed to read block 0x1a88400: -5 Sep 19 10:16:57 Svr kernel: device offline error, dev sda, sector 4005872 op 0x0:(READ) flags 0x80700 phys_seg 1 prio class 2 Sep 19 10:16:57 Svr kernel: device offline error, dev sda, sector 4005872 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2 Sep 19 10:16:57 Svr kernel: I/O error, dev loop1, sector 317712 op 0x0:(READ) flags 0x800 phys_seg 9 prio class 2 Sep 19 10:16:57 Svr kernel: device offline error, dev sda, sector 3742504 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2 Sep 19 10:16:57 Svr kernel: SQUASHFS error: Failed to read block 0x9b22804: -5 Sep 19 10:16:57 Svr kernel: I/O error, dev loop1, sector 54336 op 0x0:(READ) flags 0x800 phys_seg 13 prio class 2 Sep 19 10:16:57 Svr kernel: SQUASHFS error: Unable to read fragment cache entry [9b22804] Sep 19 10:16:57 Svr kernel: SQUASHFS error: Failed to read block 0x1a88400: -5 Sep 19 10:16:57 Svr kernel: SQUASHFS error: Unable to read fragment cache entry [9b22804] Sep 19 10:16:57 Svr kernel: SQUASHFS error: Unable to read fragment cache entry [9b22804] Sep 19 10:16:57 Svr kernel: SQUASHFS error: Unable to read page, block 9b22804, size 7828 Sep 19 10:16:57 Svr kernel: SQUASHFS error: Unable to read fragment cache entry [9b22804] Sep 19 10:16:57 Svr kernel: SQUASHFS error: Unable to read page, block 9b22804, size 7828 Sep 19 10:16:57 Svr kernel: device offline error, dev sda, sector 3742504 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2 Sep 19 10:16:57 Svr kernel: I/O error, dev loop1, sector 54336 op 0x0:(READ) flags 0x800 phys_seg 13 prio class 2 Sep 19 10:16:57 Svr kernel: SQUASHFS error: Failed to read block 0x1a88400: -5 Sep 19 10:16:57 Svr kernel: SQUASHFS error: Failed to read block 0x1a88400: -5 Sep 19 10:16:57 Svr kernel: SQUASHFS error: Failed to read block 0x1497e48: -5 Sep 19 10:16:57 Svr kernel: SQUASHFS error: Failed to read block 0x14a3d84: -5 *This continues until the hard shutdown.

Omg you are right. Thank you for that. I got my wires crossed going back and forth looking and typing.

Go to the Main tab and click the cache device. Scroll down and find the Balance button. Next to that, there is a drop down selection. Set it to 'Convert to single mode' and then click the Balance button. When it's done, you will have 2 separate drives. You will be able to verify your data is on the drive you want to keep in there. Then stop the array and remove the one you wish to swap out by changing slots from 2 to 1. Verify the one you wanted to keep is set in teh remaining cache spot. Then get your new big hdd installed, boot up, back to Main tab, and assign the hdd how you want then start the array back up.

Will need to do a New Config to remove the unwanted disk. Be sure to preserve your cache drives and assign the correct HDDs to their place... Parity and Disk1. If you mix those up, the parity build will write over your data.

I do not see anything indicating that. The top slot is connected directly to cpu like most mb are. The rest get bandwidth doled out by the chipset. If the only cards you are running are the two mentioned, you will not run into any problems. Its alot of mb for just those 2 cards.

You are right. I read it wrong in the link. Throw one of them on top. Sounds like you are using the igpu so you dont need a gpu in there. If you are forced to choose to put one of those cards in the x1 slot, I would probably do the hba provided it is full of hdds and not fast ssd.

Looks like the bottom 2 x16 slots are wired at x4. One is PCIEX 4.0 and the other is 3.0. Both are fast enough for either card you mention. If you can make them fit, put them both on bottom.

It is very peculiar that it worked for a bit then stopped. If I recall correctly, bonding is enabled by default, so disable that. Is there by chance a firmware update available for the Orbi? Try that and then resetting it to default and set it up again from scratch. Bonus points if it lets you save your current configuration somehow. Random thought too, check your times are correct on the router and server. NTP is already set most likely but I have seen off times cause strange stuff on networks before.

I understand. What I wanted to know is that if your router assigns the network information rather than assigning it on the Unraid client, would the assignment into a network (10.10.1.1/24) work as it is intended just like other random devices. If it does work, then the problem is a configuration setting on the client. AND, if it does work as a dhcp client, you could assign it the 10.10.1.2 ip address from the routers dhcp settings area. If it DOES NOT work, then the router is likely the culprit and you can look deeper there. Anyway, best of luck.

Have you tried: restarting router change router physical port set UR to DHCP vice static (verify assigned IP through router connection section status page to see if it works) testing a different cable see if any of that yields something.

I really like x710. I have the x710-DA2 but now I wish I'd spent more for the DA4.

Had a similar issue that went away when I relocated my share of iso files to the cache drive. Might be worth trying.

Move in the Main tab invokes the mover. It will move files around as per your parameters. Typically this is moving files from a cache to array. Using Unbalance Scatter function to get all your data off of a certain drive and spread among the rest will easily clear the drive you want to remove.

I've not had any benefits from using bonding. I have heard that in environments with many users gains can be observed while simultaneous users access data across the same path. Otherwise, it just adds another point of failure in a system that you likely want to remain as stable as possible.

Majority of my data being media, I just assign the disks to the shares I am feeling at the moment, set to high water, and let it ride. If something about it bothers me later on, then I use Unbalance to move things around.

That change is found under settings > docker.

The latest stable version has corrected some macvlan problems. You can update to that or alternatively, change over to ipvlan

No

Your understanding of bonding is correct. It combines multiple network connections to a single bigger one. If you run only 1 nic, it is useless. Turn it off. It is just an extra cog in the machine that can break.

Eth0 may be swapping nic choice. Quick test by swapping to the mb port. If it connects, you know what happened. Why is would be beyond me.

Assuming you have space on the rest of your array, you can move your data from disk to disk manually or if lazy like me, use the Unbalanced plugin.

Probably a few ways to go about this. 1. I would first take a pic of your main (drives) tab for reference. 2. Do something with your data on the drive about to be removed. 3. Then remove that data drive you don't want in the array anymore. 4. Then do a new config keeping your original parity drive and adding the second parity. (There are some check boxes you need to preserve data on the array drives and caches if I recall... Be sure you understand those before proceeding.) Set all your remaining data drives as the array.

This reminds me of this I just read... You using a cyberpower ups by chance? Beyond that, attach your diagnostics for better assistance.

I would just toss it in. Should take about the same time either way. I view preclear as something handy for pre staging for this scenario.

Swap the dead 2TB with the new 12TB and let it rebuild array. If you want extra precaution, turn array back on and copy your emulated data to another location first.