Squid Posted February 21, 2017 Author Share Posted February 21, 2017 1 hour ago, khronos247 said: Hi, Could I get assistance in deciphering the following Call Trace? Also what are some resources to help figure out future errors because google search was not the most helpful? Thanks! Feb 19 23:20:19 Hiveserver kernel: CPU: 0 PID: 0 Comm: swapper Not tainted 4.9.8-unRAID #1 Feb 19 23:20:19 Hiveserver kernel: Hardware name: System manufacturer System Product Name/P7H55-M LE, BIOS 0801 09/07/2012 Feb 19 23:20:19 Hiveserver kernel: ffffffff81c03d60 ffffffff813a34fa ffffffff81c03db0 ffffffff81988d29 Feb 19 23:20:19 Hiveserver kernel: ffffffff81c03da0 ffffffff8104d04c 0000034c81c03e18 ffffffff81e2801c Feb 19 23:20:19 Hiveserver kernel: ffffffff81e28058 0000000000000000 0000000000000000 0000000000014140 Feb 19 23:20:19 Hiveserver kernel: Call Trace: Feb 19 23:20:19 Hiveserver kernel: [<ffffffff813a34fa>] dump_stack+0x61/0x7e Feb 19 23:20:19 Hiveserver kernel: [<ffffffff8104d04c>] __warn+0xb8/0xd3 Feb 19 23:20:19 Hiveserver kernel: [<ffffffff8104d0f4>] warn_slowpath_fmt_taint+0x3f/0x47 Feb 19 23:20:19 Hiveserver kernel: [<ffffffff81cef157>] ? __early_ioremap+0x13e/0x1b7 Feb 19 23:20:19 Hiveserver kernel: [<ffffffff81475d16>] warn_invalid_dmar+0x70/0x7d Feb 19 23:20:19 Hiveserver kernel: [<ffffffff8167571b>] dmar_validate_one_drhd+0x8e/0xa0 Feb 19 23:20:19 Hiveserver kernel: [<ffffffff81475703>] dmar_walk_remapping_entries+0xe6/0x113 Feb 19 23:20:19 Hiveserver kernel: [<ffffffff81d01de1>] detect_intel_iommu+0x54/0xd1 Feb 19 23:20:19 Hiveserver kernel: [<ffffffff8167568d>] ? xen_swiotlb_init+0x3ab/0x3ab Feb 19 23:20:19 Hiveserver kernel: [<ffffffff81cd8c8b>] pci_iommu_alloc+0x50/0x6c Feb 19 23:20:19 Hiveserver kernel: [<ffffffff81ce3e94>] mem_init+0x17/0x9d Feb 19 23:20:19 Hiveserver kernel: [<ffffffff81ccbc98>] start_kernel+0x1d9/0x3dc Feb 19 23:20:19 Hiveserver kernel: [<ffffffff81ccb120>] ? early_idt_handler_array+0x120/0x120 Feb 19 23:20:19 Hiveserver kernel: [<ffffffff81ccb2d6>] x86_64_start_reservations+0x2a/0x2c Feb 19 23:20:19 Hiveserver kernel: [<ffffffff81ccb3be>] x86_64_start_kernel+0xe6/0xf3 Feb 19 23:20:19 Hiveserver kernel: ---[ end trace 0000000000000000 ]--- You really should create a new thread where more eyes will see it, and the entire diagnostics file as there may be other issues that led up to the call trace Quote Link to comment
Squid Posted February 25, 2017 Author Share Posted February 25, 2017 - Fixed: Don't follow symlinks to directories during an extended test 1 Quote Link to comment
Squid Posted February 25, 2017 Author Share Posted February 25, 2017 - Added in tests to detect files / folders being written to /mnt. Mainly to detect situations like this: but more thorough. Quote Link to comment
dlandon Posted February 26, 2017 Share Posted February 26, 2017 You need to ignore the /mnt/RecycleBin directory when checking the /mnt/ directories. I maintain it for the Recycle Bin plugin. It is a directory of symlinks so I can use the built in browser to browse the files from /mnt/RecycleBin. Why are you looking for directories there anyway? What is the concern with directories at /mnt/? Quote Link to comment
trurl Posted February 26, 2017 Share Posted February 26, 2017 6 hours ago, dlandon said: What is the concern with directories at /mnt/? Misconfigured dockers. Quote Link to comment
Squid Posted February 26, 2017 Author Share Posted February 26, 2017 3 minutes ago, trurl said: Misconfigured dockers. Yeah, I've seen probably a half dozen diagnostics where various apps / whatever were storing folders / files within /mnt. This check is fundamentally a bug fix for the check for disks being able to be written to. I've got no problem making an exception for Recycle Bin (already there for UD) Quote Link to comment
dlandon Posted February 26, 2017 Share Posted February 26, 2017 1 minute ago, Squid said: Yeah, I've seen probably a half dozen diagnostics where various apps / whatever were storing folders / files within /mnt. This check is fundamentally a bug fix for the check for disks being able to be written to. I've got no problem making an exception for Recycle Bin (already there for UD) Thanks that works. I figured you picked up on UD and excluded the /mnt/disks. Already had a poster think there was a disaster pending because of the RecycleBin folder message in FCP. Why would apps store folders/files in /mnt/? I really didn't want to do it for the Recycle Bin, but could not get the browse to work any other way because the browser branches from /mnt/. Quote Link to comment
Squid Posted February 26, 2017 Author Share Posted February 26, 2017 3 minutes ago, dlandon said: Why would apps store folders/files in /mnt/? One way is misconfiguring /config to be mapped to /mnt Quote Link to comment
trurl Posted February 26, 2017 Share Posted February 26, 2017 15 minutes ago, dlandon said: Why would apps store folders/files in /mnt/? They shouldn't, but some make mistakes, and some don't really understand volume mapping. Quote Link to comment
Squid Posted February 26, 2017 Author Share Posted February 26, 2017 And like every other test, FCP highlights what tends to trip people up. If you know what you're doing, then there's no reason why you shouldn't put stuff within /mnt (and then ignore the error) Quote Link to comment
Squid Posted February 27, 2017 Author Share Posted February 27, 2017 7 hours ago, dlandon said: You need to ignore the /mnt/RecycleBin directory when checking the /mnt/ directories Updated Quote Link to comment
dlandon Posted February 27, 2017 Share Posted February 27, 2017 27 minutes ago, Squid said: Updated Thank you. Quote Link to comment
Squid Posted March 1, 2017 Author Share Posted March 1, 2017 - Added: Output of mcelog is now logged when an mce error is found - Added: Warnings when irq xx: nobody cared errors are found. Output of cat /proc/interrupts is logged - Added: Check if maximum number of inotify watches has been exhausted. When an irq nobody cared is found and a call trace is associated with it, then a warning (irq nobody cared) and an error (call trace) happen. Still not at the point where we can determine what exactly caused the call trace and then ignore them on a case-by-case basis, as that's a big revamp of the testing procedure, but in the interim at least the output of cat /proc/interrupts will help to determine whether or not to ignore the call trace for the time being. Quote Link to comment
etsi Posted March 5, 2017 Share Posted March 5, 2017 (edited) What the following means? I was running the trial and upgraded the license. Multiple registration keys found - While unRaid will operate with multiple .key (registration files) within the config folder on the flash drive (ie: For simplicity sake you have multiple valid registrations and are storing them all within the config folder, you will run into problems if you ever need to transfer one of the registrations to another USB stick, as unRaid will not know which registration file to transfer, and the incorrect registration may get blacklisted. You should investigate and determine which key belongs to which USB stick and only have that particular key on that stick Edited March 5, 2017 by etsi fixed: deleted /boot/config/Trial.key Quote Link to comment
trurl Posted March 5, 2017 Share Posted March 5, 2017 4 hours ago, etsi said: What the following means? I was running the trial and upgraded the license. Multiple registration keys found - While unRaid will operate with multiple .key (registration files) within the config folder on the flash drive (ie: For simplicity sake you have multiple valid registrations and are storing them all within the config folder, you will run into problems if you ever need to transfer one of the registrations to another USB stick, as unRaid will not know which registration file to transfer, and the incorrect registration may get blacklisted. You should investigate and determine which key belongs to which USB stick and only have that particular key on that stick Delete your trial key. Not sure there is any real issue in this case, but there have been cases where things got confused due to having more than one key on flash. 1 Quote Link to comment
Spritzup Posted March 5, 2017 Share Posted March 5, 2017 So I'm seeing the following error with the plugin "Possibly out of inotify watches". However, I don't have any of the 'problem' plugins installed... the only thing that I can think of causing the problem would be Plex monitoring for file changes, as that seems to use inotify. Here is a list of the plugins I have installed --> CA Auto Update Applications CA Backup/Restore Appdata CA Cleanup Appdata Community Applications Dynamix Cache Directories Dynamix Local Master Dynamix SSD Trim Dynamix System Statistics Dynamix WebGui Fix Common Problems IPMI Support Nerd Tools Open Files Preclear Disks Recycle Bin Tips and Tweaks Unassigned Devices unBalance unRaid Server OS Thanks in advance! ~Spritz Quote Link to comment
JustinAiken Posted March 9, 2017 Share Posted March 9, 2017 I have a massive amount of "The following files exist within the same folder on more than one disk."'s... What's the best way to fix that? Do I need to just manually `stat` the files on both disks and manually delete the older/smaller one? Or is there an automated way to fix? Quote Link to comment
RobJ Posted March 10, 2017 Share Posted March 10, 2017 8 hours ago, JustinAiken said: I have a massive amount of "The following files exist within the same folder on more than one disk."'s... What's the best way to fix that? Do I need to just manually `stat` the files on both disks and manually delete the older/smaller one? Or is there an automated way to fix? Try unRAIDFindDuplicates.sh, I think it should help. Otherwise, I would use a double pane file manager, like MC or Krusader or Dolphin, to compare folders and delete the duplicated folders and files. Quote Link to comment
hokeysmoke Posted March 17, 2017 Share Posted March 17, 2017 (edited) False positive on a folder with the difference being a trailing zero, but detected as a case difference: The following directories exist with similar names, only differing by the 'case' which will play havoc with Windows / SMB access. Windows does NOT support folder names only differing by their case and strange results will happen should you attempt to manipulate the folders or files/mnt/user/Documents/Document Backup/Nicholas-PC/AppData/Roaming/.minecraft/versions/1.1/mnt/user/Documents/Document Backup/Nicholas-PC/AppData/Roaming/.minecraft/versions/1.10 Edited March 17, 2017 by hokeysmoke clarification Quote Link to comment
Squid Posted March 17, 2017 Author Share Posted March 17, 2017 (edited) 6 hours ago, hokeysmoke said: False positive on a folder with the difference being a trailing zero, but detected as a case difference: The following directories exist with similar names, only differing by the 'case' which will play havoc with Windows / SMB access. Windows does NOT support folder names only differing by their case and strange results will happen should you attempt to manipulate the folders or files/mnt/user/Documents/Document Backup/Nicholas-PC/AppData/Roaming/.minecraft/versions/1.1/mnt/user/Documents/Document Backup/Nicholas-PC/AppData/Roaming/.minecraft/versions/1.10 Found and fixed. It wasn't trailing zero's, but any folder name that could be considered a number. IE: 000000 and 0 would be considered the same, along with 1 and 000001 Will package up the fix and release after work. Edited March 17, 2017 by Squid Quote Link to comment
hokeysmoke Posted March 18, 2017 Share Posted March 18, 2017 Thanks! Ran it again and confirmed fixed. Quote Link to comment
itimpi Posted March 19, 2017 Share Posted March 19, 2017 I noticed that I was getting a warning about inotify-tools being installed but no longer needed as I am on 6.3.2 and it is now built it. However when I go to the nerd pack settings the inotify-tools package is not shown to enable the issue to be fixed. Manually editing the underlying .cfg file for nerd pack shows that the option is set to "yes" so I can clear it by changing that to "no". However I should be able to do this via the GUI I would have thought? Not sure if the fix needs to be to nerd pack plugin or the FCP plugin? Quote Link to comment
Squid Posted March 19, 2017 Author Share Posted March 19, 2017 (edited) 4 hours ago, itimpi said: I noticed that I was getting a warning about inotify-tools being installed but no longer needed as I am on 6.3.2 and it is now built it. However when I go to the nerd pack settings the inotify-tools package is not shown to enable the issue to be fixed. Manually editing the underlying .cfg file for nerd pack shows that the option is set to "yes" so I can clear it by changing that to "no". However I should be able to do this via the GUI I would have thought? Not sure if the fix needs to be to nerd pack plugin or the FCP plugin? But it probably is now a waste of a test since even with the .cfg file set to install it, dmacias has removed the package from nerd pack. Next update.... Edited March 19, 2017 by Squid Quote Link to comment
dlandon Posted March 22, 2017 Share Posted March 22, 2017 Got a minor typo in one of the FCP log messages: Mar 20 18:39:29 Tower root: Fix Common Problems: Error: Probable 32 Big package ntfs-3g-2010.3.6-i486-1.txz found on the flash drive in the packages folder I think you meant to say "32 Bit package..." not "32 Big package...". I'm not sure a "Big" package is necessarily a problem. Quote Link to comment
tunetyme Posted April 6, 2017 Share Posted April 6, 2017 I have recently upgrades my system from 4.7 to 6.3.2 and have upgraded several disks from 2 to 4TB. Current configuration is: 6 data XFS 1 cache BTRFS 1 parity 2 empty slots disks 7 & 8 (used for preclearing, copying etc. 2 shares P and S I went through the conversion process from rfs to xfs. In that process there was a recommendation to have both include in each plus exclude slots 7 & 8 (disk). I have an error for both shares saying I am including and excluding and to correct this. Both share indicate that neither share is excluding any disks. However it also only shows 6 disks not 8. How can I correct this? Quote Link to comment
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.