partyhat

So I have rebooted again and run xfs-repair. It seems to be running okay for now so I will monitor the disk.

I upgraded from 6.9.2 to 6.10.3 just this afternoon. The server was working okay for about 6 hours, but then my 18TB disk has dropped offline twice in different ways - first it went to unassigned devices and clicking the disk log showed the following error messages: Aug 21 16:37:24 Tower wsdd2[7429]: starting. Aug 21 23:06:16 Tower kernel: sd 4:0:0:0: [sdd] Synchronizing SCSI cache Aug 21 23:06:16 Tower kernel: sd 4:0:0:0: [sdd] Synchronize Cache(10) failed: Result: hostbyte=0x04 driverbyte=DRIVER_OK Aug 21 23:06:16 Tower kernel: sd 4:0:0:0: [sdd] Stopping disk Aug 21 23:06:16 Tower kernel: sd 4:0:0:0: [sdd] Start/Stop Unit failed: Result: hostbyte=0x04 driverbyte=DRIVER_OK Aug 21 23:07:21 Tower emhttpd: error: hotplug_devices, 1719: No such file or directory (2): Error: tagged device WDC_WD180EDGZ-11B2DA0_2JGPE3MG was (sdd) is now (sdg) Aug 21 23:45:55 Tower kernel: XFS (sdd1): log I/O error -5 Aug 21 23:45:55 Tower kernel: XFS (sdd1): Log I/O Error (0x2) detected at xlog_ioend_work+0x50/0x6e [xfs] (fs/xfs/xfs_log.c:1377). Shutting down filesystem. Aug 21 23:45:55 Tower kernel: XFS (sdd1): Please unmount the filesystem and rectify the problem(s) I took diagnostics which are attached to this post. I googled and found a link to the forum that suggested to reboot the server, which I've done and it came back online okay. In the next 10 minutes after startup, it dropped off again and now shows under 'historical devices'. I took more diagnostics which are attached too. The drive still appears in the cache pool area (it's its own pool) but when I click to view its contents, rather than listing my folders it looks like it lists a linux filesystem, per the attached image. Thought this was very odd. I hope I don't have to rollback - even if that's a thing. Any help would be appreciated, thank you. Mike tower-diagnostics-20220821-2351.zip tower-diagnostics-20220822-0007.zip

Thanks I ran the Xfs repair and it seemed to work for now. I have done the same thing a couple times recently Too so will see how it goes. I did open the box and spray all the connections with contact cleaner so that may have helped too.

Hi, had this issue on and off a couple weeks now. Seems when downloading or moving files the user shares all disappear and if I try view files from the Main tab it says “No listing: too many files”. I had searched the forum and found a post suggesting to turn on disk shares and disable hard links. I did this but the error persists. Rebooting seems to fix things only temporarily then it will happen again. Looked at a few topics and none seem to have satisfactory resolution, is this a known bug? Per this thread (https://forums.unraid.net/topic/74731-solved-all-user-shares-have-disappeared-from-unraid-gui/page/2/ I reformed ls -ail /mnt command and got the attached screenshot with a bunch of ? Text. Then I rebooted and ran command again before starting the array and after starting it and it looks better as per the next screenshot. Ran the chmod 777 /mnt/user as the thread suggested. Shares still disappeared after a while Diagnostics attached. tower-diagnostics-20220306-1209.zip

Ok thanks. I’m bidding on a replacement main board on eBay. If I got the same motherboard and swapped it, would that work straight up or is there likely to be something like serial numbers/hardware ids that mean I’ll need to setup a new unraid USB key and rebuild the array with the same disks?

Yes I should have said, the other drive sometimes comes up as 'unmountable disk present'. What I do is bring the array online in maintenance mode, then run the xfs repair on the disk and delete the log. Then that one comes online again fine. The thing is looking at the smart logs, UDMA CRC error count looks to be zero, unless I'm missing something and I thought that would indicate the backplane issue?

I've had this happen a few times now, where the parity drive (8TB HGST - disk sdf) gets a red x next to it with read errors. I'd run SMART tests and they would usually come back fine. I had then stopped the array, set the parity slot to 'no device'. Rebooted the server, added the parity drive back to the slot and started the array, a rebuild would occur and it would be okay again for maybe a month. This happened at least three times now so it is getting annoying. The disk currently shows the red x under the 'array devices' section with drive ID SDF. It's also concurrently under the 'unassigned devices' section with drive ID SDG. The hardware is an HP N36L Microserver, so about 11 years old and getting a bit aged. I just wonder if it's the backplane or cable for this specific drive, or in actual fact the disk itself is dieing.. Diagnostics attached. If it's clearly not the disk I could try reassigning disks to different slots to check the slot, but I don't want to do that yet as I'm not confident how to do it. Any help would be much appreciated, thanks! tower-diagnostics-20220217-1331.zip

I guess the point I’m trying to make now is what is the point of the “sending email address” field if it only comes through the email system as a way to pointlessly reply to automated emails? It’s certainly not the sender of the notifications as the help description suggests.

Weird, does your email header show it’s sent from the value in the “sending email address” field? I’m on 6.9.2 by the way.

Yes, using an Application Password. I've just created another application password and the same behaviour occurs. I've just routed it from an old/unused gmail account to my primary now, so at least it won't fill my Sent items label

Fair enough, if that’s how they’ve developed it to work.. I think the “help” info should be refined though, as it isn’t really doing what it suggests. I figured it should spoof the sent-from info the same way it spoofed the reply-to info. Replying to a notification at that address would obviously never be done, or helpful since it’s an obviously made up email address. For now I guess I’ll just route it via another gmail account, though it’s not ideal. Thanks for your help, and maybe the dev team could look at this further.

A rule to do what? Isn’t this fixable with unraid, as it appears to be a bug that it doesn’t send from the address it should be?

I have Notifications configured as below but it fills up my "sent" items in Gmail with the daily notifications which is frustrating when trying to look at my sent items. I changed the sending email address but I still get it showing up in "Sent" and it still appears in the "From" field, with the sending address only shown for reply-to, as in the third screenshot below. How to fix this? Settings: "Sent" folder: Received email:

I have this configured but it fills up my "sent" items in Gmail with the daily notifications. I changed the sending email address but I still get it showing up in "Sent" and it still appears in the "From" field, with the sending address only shown for reply-to, as in the third screenshot below. How to fix this? Settings: Sent folder: Received email:

Yes there's a destination set and it's adjacent to the appdata and libvirt backup folders. I should mention the Unraid USB key that was in use when I setup CA B+R (last year), failed the day before the last successful backup. Seems a weird coincidence, but I think the current backup is of the replacement USB key. Perhaps I just need to remove the app and reinstall? I didn't want to do that incase it messed up anything though. I noticed above you said the flash backup will be deprecated? Maybe I should be looking for another solution to backup flash if this is the case. If so, any recommendations?