madburg

Members
  • Posts

    1277
  • Joined

  • Last visited

Everything posted by madburg

  1. Yes, but my personal take to this is I rather preclear the drive ( you have options to make it skip a step to speed things up and only one pass) and add it to the array, versus unassign parity and issue a new config, and the smallest chance that lightning strikes you would be sorry. But yes one could do as you stated in your steps.
  2. r.e. "... parity is written with all the 0/1 from this drive ..." ==> NO. When a drive is added to your protected array, NO parity computations are done ... Stop right there, no computations are done: Question 1) because it thinks we have all zero's on the drive, correct? Question 2) something has to change, as we added an additional drive for which it must now track and maintain parity for, no? Please read, I faked the signature on the drive i added, unRAID did not zero (clear) this drive out and just formated the files system, only. Parity does no computation, right, so I add three files to this new drive, and this drive fails a week later, I add a new one (allowing it to clear and format it), wouldn't parity rebuilt these 3 files on this new drive? if not what would it do? (I don't know, I am asking.)
  3. If it was written to at all, the preclear is no longer valid... Agreed for any drive one wishes to add as a data drive, unRAID doen not have this check for a drive you assign to as parity, correct? So that answers my question, thanks, it will start to clear the drive (thus array would not be available, if you do this with a drive that has not been currently precleared and not messed with). Off-topic, sorry to the OP ahead of time. This part I don't get, say a drive is not zero'ed out and a signature is faked, parity is written with all the 0/1 from this drive. As data is added to the drive it changes varies bits to 0/1's, so lets pretend I added 3 files only, and this drive fails I add a new one, wouldn't parity rebuilt these 3 files and whatever else (0/1's) it had from faking it out to begin with? thats what I am not completely understanding...
  4. unRAID would have lost track of the fact that drive had been pre-cleared before using it is a parity drive. That is quite understood, thanks though (never hurts should someone else be reading this thread). However, unRAID never really knew it was precleared to begin with though, as there is no signature check for a parity drive. Preclearing a drive that will be used for the parity drive assignment is just something (recommended) someone do to make sure the drive is OK to begin with (HW wise). So you are stating that removing the parity hard drive and assigning it as a data disk in the array would make unRAID start clearing the drive, not just formating it, is that correct? ---------- unRAID step 1 is no signature found on HD, clear the drive. (lengthy process array not available) unRAID step 2 format the drive. (pretty quick, array is available) ----------
  5. What assignment DID you have these drives in before you started this process? e.g. Disk1, Disk3, Disk5, Disk7 (Disk2,4,6,8-24 unassigned, no parity) (its understood one of the 5 was allocated for parity.) P.S. i have a feeling you removed parity and disk#3 at the same time, but don't want to jump to conclusions.
  6. Several things. 1) Your Bios is running the original firmware (2007), I recommend you follow the mft. directions very closely and update to the latest Version 0902 (2009), please verifiy the link is for your board. http://www.asus.com/Motherboards/P5EVM_DO/#support_Download_5 2) Syslog shows: Jul 10 20:41:14 Tower kernel: 97MB HIGHMEM available. If I am not mistaken you have 1GB of ram, this is odd (update the bios) 3) syslog shows: Jul 10 20:41:14 Tower kernel: ACPI Warning: 0x00000400-0x0000041f SystemIO conflicts with Region \SMRG 1 (20130117/utaddress-251) most likely kernel issue, but until you update to the latest bios, this would not be far to look into. 4) Jul 10 20:41:16 Tower emhttp: get_config_idx: fopen /boot/config/shares/0DAY.cfg: No such file or directory - assigning defaults Jul 10 20:41:16 Tower emhttp: get_config_idx: fopen /boot/config/shares/BACKUP.cfg: No such file or directory - assigning defaults Jul 10 20:41:16 Tower emhttp: get_config_idx: fopen /boot/config/shares/EBOOKS.cfg: No such file or directory - assigning defaults Jul 10 20:41:16 Tower emhttp: get_config_idx: fopen /boot/config/shares/GAMES.cfg: No such file or directory - assigning defaults Jul 10 20:41:16 Tower emhttp: get_config_idx: fopen /boot/config/shares/ISEC.cfg: No such file or directory - assigning defaults Jul 10 20:41:16 Tower emhttp: get_config_idx: fopen /boot/config/shares/ISOS.cfg: No such file or directory - assigning defaults Jul 10 20:41:16 Tower emhttp: get_config_idx: fopen /boot/config/shares/LEARN AND MASTER.cfg: No such file or directory - assigning defaults Jul 10 20:41:16 Tower emhttp: get_config_idx: fopen /boot/config/shares/MP3.cfg: No such file or directory - assigning defaults Jul 10 20:41:16 Tower emhttp: get_config_idx: fopen /boot/config/shares/Movies.cfg: No such file or directory - assigning defaults Jul 10 20:41:16 Tower emhttp: get_config_idx: fopen /boot/config/shares/MoviesDB.cfg: No such file or directory - assigning defaults Jul 10 20:41:16 Tower emhttp: get_config_idx: fopen /boot/config/shares/P90X.cfg: No such file or directory - assigning defaults Jul 10 20:41:16 Tower emhttp: get_config_idx: fopen /boot/config/shares/SERIES.cfg: No such file or directory - assigning defaults This is not good, can you please browse to your flash drive, config/shares folder and IF there are any files there, zip and attach for review.
  7. Can you run the following test, change the "Allocation method:" to "most free" for the movies share and re-test your original "copy both to a smb share and directly to the device". You can change it back afterwards.
  8. If your original parity drive was precleared originally to make sure all was well with it and then was added as a parity. Now as hunter69 mentioned you wish to remove it to upgrade the parity size and reuse this drive in the array; what happens IF you took that old parity drive and didn't preclear it (again) and added it directly to the array as another addition (NEW) disk? Would unRAID start a clearing of the drive or just give you the option to format it quickly and add it in?
  9. In the latest releases, the safemode flag causes the standard go file to be executed, because there are a few users that require certain statements in their go file for their array to work correctly. This means that while plugins and addons in the extra folder are not loaded, addons like UnMENU are still loaded, including any auto-install UnMENU addons. I would prefer a little different startup. Ask users who require special commands to place them in both their regular go file and in the new go-safe file, which will only be executed in safe mode. Then Tom would change his startup logic to: if unraidsafemode flags if go-safe exists execute go-safe else load emhttp else load plugins etc... Easy to implement, easy for Tom, easy to explain, easier to support because lower chance of syntax errors if users are trying to insert safemode conditional statements... That was my understanding of how I thought it works, what is the go-safe file if not for the " few users that require certain statements in their go file for their array to work correctly". Isn't the go file skipped and go-safe run instead?
  10. e.g. User Share = "TV SHows" / Disks Included = disk1-4 Disk1 Top level directory = "TV SHows" Disk2 Top level directory = "TV SHows" (user.org.netatalk.supports-eas.xxx) Disk3 Top level directory = "TV SHows" Disk4 Top level directory = "TV SHows" So you would run: getfattr –d /mnt/disk1/TV\ Shows/ getfattr –d /mnt/disk2/TV\ Shows/ getfattr –d /mnt/disk3/TV\ Shows/ getfattr –d /mnt/disk4/TV\ Shows/ You would notice that when you got to disk #2 the command outputted you had an "user.org.netatalk.supports-eas.xxx" So now you would run: /boot/rmattrs "/mnt/disk2/TV Shows" Now verify the apple eas extended attribute is now gone: getfattr –d /mnt/disk2/TV\ Shows/ Don't run against User Share nor a Disk itself, run "getfattr –d" against the top level directory on EACH disk, to see whether it contains a eas extended attribute. If it finds one or more, run Tom's script to have them removed them. The script is actually creating a new directory (e.g. TV Shows~temp) on the disk, moving all the data from your current (e.g. TV Shows) directory you selected to run against and then deletes the original (e.g. TV Shows) directory (Because the apple eas extended attribute is tied to the top level directory). Then re-names the temp (e.g. TV Shows~temp) directory to the original name (e.g. TV Shows). So you have to follow the directions if you don't want to lose data; hence why Tom want you to go one by one and stop immediately if something went wrong. Hope I don't get a shellacking but you could speed up the process of FINDING if you have apple eas extended attributes by running getfattr –dR /mnt/user (if you have a cache drive) getfattr –dR /mnt/user0 (if you don't have a cache drive) So it would scan all in one shot for you and should any come back you should note the DISK & Top Level Directory it references the apple eas extend attribute resides on, so you know which disk and top level directory to run the "rmattrs" against. Note as an example the following is not an apple eas extended attribute, it is a file extended attribute and IS correct: uninstall_flash_player_osx.dmg user.com.apple.metadata:kMDItemDownloadedDate=0sYnBsaXN0MDChATNBtRZGToOIIggKAAAAAAAAAQEAAAAAAAAAAgAAAAAAAAAAAAAAAAAAABM= user.com.apple.metadata:kMDItemWhereFroms=0sYnBsaXN0MDCiAQJfEFdodHRwOi8vZnBkb3dubG9hZC5tYWNyb21lZGlhLmNvbS9nZXQvZmxhc2hwbGF5ZXIvY3VycmVudC91bmluc3RhbGxfZmxhc2hfcGxheWVyX29zeC5kbWdfEElodHRwOi8vaGVscHguYWRvYmUuY29tL2ZsYXNoLXBsYXllci9rYi91bmluc3RhbGwtZmxhc2gtcGxheWVyLW1hYy1vcy5odG1sCAtlAAAAAAAAAQEAAAAAAAAAAwAAAAAAAAAAAAAAAAAAALE= user.com.apple.quarantine="0001;4f660ece;Safari;605F0A3D-7CE8-4202-A649-C2A42459336C|com.apple.Safari"
  11. Going over a few things in RC16c I ran into the following while checking the syslog: Jul 12 16:04:40 Tower logger: +============================================================================== Jul 12 16:04:40 Tower logger: | Installing new package /mnt/vmdk/packages/open_vm_tools-2013.04.16.1098359_unRaid5.0rc16c-i686-24Zeron.tgz Jul 12 16:04:40 Tower logger: +============================================================================== Jul 12 16:04:40 Tower logger: Jul 12 16:04:40 Tower logger: Verifying package open_vm_tools-2013.04.16.1098359_unRaid5.0rc16c-i686-24Zeron.tgz. Jul 12 16:04:40 Tower logger: Installing package open_vm_tools-2013.04.16.1098359_unRaid5.0rc16c-i686-24Zeron.tgz: Jul 12 16:04:40 Tower logger: PACKAGE DESCRIPTION: Jul 12 16:04:40 Tower logger: Executing install script for open_vm_tools-2013.04.16.1098359_unRaid5.0rc16c-i686-24Zeron.tgz. Jul 12 16:04:40 Tower logger: ^[[1;32m _____ _____ _____ _____ _^[[0m Jul 12 16:04:40 Tower logger: ^[[1;32m | |___ ___ ___ | | | | |_ _|___ ___| |___^[[0m Jul 12 16:04:40 Tower logger: ^[[1;32m | | | . | -_| | | | | | | | | | | . | . | |_ -|^[[0m Jul 12 16:04:40 Tower logger: ^[[1;32m |_____| _|___|_|_| \___/|_|_|_| |_| |___|___|_|___|^[[0m Jul 12 16:04:40 Tower logger: ^[[1;32m |_| ^[[1;34mv. 2013.04.16-1098359 for unRAID 5.0-rc16c^[[0m Jul 12 16:04:40 Tower logger: Jul 12 16:04:40 Tower logger: ^[[0;33mLoading VMWare modules...^[[0m Jul 12 16:04:40 Tower logger: ^[[0;33m modprobe vmw_vmci^[[0m Jul 12 16:04:40 Tower logger: ^[[0;33m modprobe vmhgfs^[[0m Jul 12 16:04:40 Tower logger: ^[[0;33mStarting VMWare Tools Daemon.^[[0m Jul 12 16:04:40 Tower kernel: vmw_vmci 0000:00:07.7: Found VMCI PCI device at 0x11080, irq 16 Jul 12 16:04:40 Tower kernel: vmw_vmci 0000:00:07.7: Using capabilities 0xc Jul 12 16:04:40 Tower kernel: vmw_vmci 0000:00:07.7: irq 76 for MSI/MSI-X Jul 12 16:04:40 Tower kernel: vmw_vmci 0000:00:07.7: irq 77 for MSI/MSI-X Jul 12 16:04:40 Tower kernel: Guest personality initialized and is active Jul 12 16:04:40 Tower kernel: VMCI host device registered (name=vmci, major=10, minor=59) Jul 12 16:04:40 Tower kernel: Initialized host personality Jul 12 16:04:40 Tower vmsvc[1601]: [ warning] [GLib-GObject] invalid (NULL) pointer instance Jul 12 16:04:40 Tower vmsvc[1601]: [critical] [GLib-GObject] g_signal_emit_by_name: assertion `G_TYPE_CHECK_INSTANCE (instance)' failed Jul 12 16:04:40 Tower logger: ^[[0;33mAttempting to mount /mnt/hgfs^[[0m Jul 12 16:04:40 Tower logger: Package open_vm_tools-2013.04.16.1098359_unRaid5.0rc16c-i686-24Zeron.tgz installed. Jul 12 16:04:40 Tower logger: Jul 12 16:04:40 Tower logger: Jul 12 16:04:40 Tower logger: success Jul 12 16:04:40 Tower logger: plugin successfully installed I don't recall seeing the warning and critical lines before, something that may need looking into. I also noticed, and maybe its an updated kernel thing, but thought I should mention should it be vmtools related and if anyone less see's this behaviour. There are times when the vsphere client shows Zero usage of active guest memory, where in the past RC's and vmtools loaded there was at least some usage displayed never zero. Screenshot attached to lessen any confusion of what I am referring to. THis may not be an issue but a good, thing, just not sure if an OS that's running actually is using nothing...
  12. To all testing if a drive is truly spun up or down here. This command is used to obtain spun up/down status: hdparm -C /dev/sdX e.g. hdparm -C /dev/sdb If drive is spun up you should see this: /dev/sdb: drive state is: active/idle If drive is spun down, you should see this: /dev/sdb: drive state is: standby Based on the output from hdparm you will know if the gui you are using (unRAID Main, unMenu, SF) is or is not displaying properly for some reason. The following command (lsof "list open files") can be used to see if anything is open on the suspect disk as well: lsof | grep mnt
  13. @EMKO, please post your full syslog, when you have a moment. Download hdparm and smartmontools updates from this posting : http://lime-technology.com/forum/index.php?topic=28382.msg252874#msg252874 (original poster of these pulled them unfortunitly, long story; or I would have had you reference his). Update your system with those two packages, re-try your test.
  14. It's mostly one component on its way out. Do you have any motherboard to swap out temporarily with? Or another PSU, it doesn't have to fit your case, just hanging of to the side is fine. This way u could rule out what went south on you. Borrow a friends PSU, bring ur case to their house, I am sure you could think of something. So u know what you truly need to buy.
  15. None of those should exist. That is a damaged/corrupted one, running mover (provided u have a cache drive) or rync with extended attribute support script will choke to the movie user share. These two got written and should have been immediately removed, weren't and are read each time only confusing matters. Tom's script will remove all three, none will come back. His script may have an issue with spaces in a share name, so I would hold until he comments back. In your case "Movies" does not contain a space in the name and can be run again the "Movie" directory for EACH disk it resides on.
  16. Yes, you don't want ANY entries with "user.org.netatalk.supports-eas.xxxx" to come back. Its important to note (Tom should add) depending on how many files you have, is how long it will take to search for extended attributes, and if it finds them, the move and move back will depend on how many files as well (if I am not mistaken).
  17. hmm that should not be the case... Which drives is this?
  18. Anytime I would sticky this and maybe put a link in the 5.0 docs your working on completing.
  19. ==> No you can't, and NO but's <=== Here we go again, you left out how long you have been doing this for... Look its real simple, your comment help him how much ZERO, no one is saying its an unRAID issue, he's having a problem help or don't comment just to post. If some info is missing ask him for it, he was asked to remove ram he did, but it didn't help him. Move his post elsewhere and let someone help him.
  20. Did you preclear the drives using JoeL.'s script before adding them to the array?
  21. I'm a Windows guy as well ... but I do know that Linux does a good job of managing physical address extension (PAE) on systems that support it. PAE is how Windows server OS's can manage far more than 4GB of RAM on 32-bit servers [both Server 2003 and Server 2008 support up to 64GB in the 32-bit versions]. Microsoft intentionally did not include PAE support in their client OS's ... although there are several add-ons that provided ways to use more RAM ... notably a couple of commercial RAMDisk utilities that would use PAE to implement RAMDisks with memory above 4GB. Bottom line: The Linux kernel supports PAE, so that's why folks often install > 4GB ==>You don't know Windows to well either.<== http://msdn.microsoft.com/en-us/library/windows/desktop/aa366796(v=vs.85).aspx PAE is HW dependent as well, not just the OS supporting it.
  22. Dude unfortunately you're having some hardware issues (cmos, cmos battery, etc..?) it may be a coincidence with the storm you had, or maybe so. Recheck/seat all your connection, psu to MB, sata/sas cables etc... worth a shot.
  23. correction #2 = ESXi utilizing RDM, issues with spindown as VMDirectPath works just fine. Nothing is delaying final, Tom knows just how much or little before he stamps final. We can all ask and point things out but he decides.
  24. Right, cool. So you need to find whats going on with your fan(s) RPMs, etc... then. Good luck.