Page 2 of 3 FirstFirst 123 LastLast
Results 11 to 20 of 28
  1. #11
    Untangle Ninja
    Join Date
    Jan 2011
    Posts
    1,325

    Default

    yep, mine is working with md raid; still no idea why it didn't work the first time. it booted up fine and /proc/mdstat shows all is well.

    even though I used the 9.0build2 cd there were still some updates to be installed right away. those went in without incident, so now I'm pre-loading the lite package (though I'm not sure why... the client's internet connection is 10x faster than mine!). this box will have WAN balancer/failover too, but I'll probably load that after the client plugs it in at their site. all that's left now is to pre-configure the WAN interface for their ISP and ship it off.

    as an aside, I just love these boards with ipmi management. I've yet to connect a keyboard, mouse, monitor or even a CD drive to this box, nor have I once physically pushed the power button. I've done absolutely everything over the network... didn't even have to burn the Untangle ISO to a disk, just 'inserted' the iso into the virtual cd drive. It's a lot like installing a VM, except there's real hardware when it's done. It's not new to me, I've spec'd all my regular servers with IPMI for the past year or so, but it's still amazing every time.

  2. #12
    Untangle Ninja
    Join Date
    Jan 2011
    Posts
    1,325

    Default

    Quote Originally Posted by sky-knight View Post
    Also, why are you mirroring your swap? Do you want the box to croak?
    if you don't mirror swap and a drive dies, the box hard locks. you have to mirror swap. there used to be some concerns over stability with swap on md, but not anymore (that I know of).

    edit: regarding where I put swap, I just duplicated the standard Untangle partition layout - they put 80gb of / first, then 1.7gb of swap, then /data for the rest. Since things weren't working, I wanted to eliminate as many other variables as possible.
    Last edited by johnsonx42; 06-14-2011 at 04:55 PM.

  3. #13
    Untangle Ninja
    Join Date
    Jan 2011
    Posts
    1,325

    Default

    by the way, thanks also for taking some time to do a test installation, and thanks for the chat about these matters.

  4. #14
    Untangle Ninja sky-knight's Avatar
    Join Date
    Apr 2008
    Location
    Phoenix, AZ
    Posts
    26,230

    Default

    That's new to me, I just put swap on both drives and if one goes... oh well. Granted, I haven't tested that theory in a while... and given these abilities in the stock installer Jim and I are discussing the possibility of using this in a production server.

    I've used camstudio to record what I've done so far... I'm cutting it up into some semblance of sanity before I send it into youtube.
    Last edited by sky-knight; 06-14-2011 at 04:59 PM.
    Rob Sandling, BS:SWE, MCP
    NexgenAppliances.com
    Phone: 866-794-8879 x201
    Email: support@nexgenappliances.com

  5. #15
    Untangle Ninja
    Join Date
    Jan 2011
    Posts
    1,325

    Default

    my understanding is that the kernel will automatically spread swap across all drives with swap partitions. however if any of that swap space is in use when a drive dies, the system can't survive the loss (obviously given what swap space holds, it couldn't survive if any of it just disappears). in many cases swap space never actually gets used, so it might survive the loss of a swap partition, but only due to luck.

  6. #16
    Untangle Ninja
    Join Date
    Jan 2011
    Posts
    1,325

    Default

    actually I should have said if a swap partition disappears, the kernel will panic. same end result, but more correct terminology.

    I will be interested to see your video. I still don't know why mine failed the first time with md raid (I tried 3 or 4 times with dmraid), but worked the second time with no significant change that I can think of. perhaps watching someone else do it will help me see what I might have done. at the end of the day it will probably be the same old story - sometimes things just don't work the first time around, only The Shadow knows why.
    Last edited by johnsonx42; 06-14-2011 at 05:19 PM.

  7. #17
    Untangle Ninja sky-knight's Avatar
    Join Date
    Apr 2008
    Location
    Phoenix, AZ
    Posts
    26,230

    Default

    http://www.youtube.com/watch?v=HCF91Q7G8yw

    Yeah, this needs testing... the kind of testing you can only get when you yank a drive out of a running system.

    And your theory about the swap makes sense logically, you may be right.
    Rob Sandling, BS:SWE, MCP
    NexgenAppliances.com
    Phone: 866-794-8879 x201
    Email: support@nexgenappliances.com

  8. #18
    Untangle Ninja
    Join Date
    Jan 2011
    Posts
    1,325

    Default

    nice video. my mdadm --detail /dev/md0 is exactly the same, down to the same block count:
    Code:
    ~ # mdadm --detail /dev/md0                                   [root @ hostname]
    /dev/md0:
            Version : 00.90
      Creation Time : Tue Jun 14 05:25:54 2011
         Raid Level : raid1
         Array Size : 78123968 (74.50 GiB 80.00 GB)
      Used Dev Size : 78123968 (74.50 GiB 80.00 GB)
       Raid Devices : 2
      Total Devices : 2
    Preferred Minor : 0
        Persistence : Superblock is persistent
    
        Update Time : Tue Jun 14 10:39:48 2011
              State : clean
     Active Devices : 2
    Working Devices : 2
     Failed Devices : 0
      Spare Devices : 0
    
               UUID : b1b17642:49d02f9d:24a24a34:3b420ac5
             Events : 0.6
    
        Number   Major   Minor   RaidDevice State
           0       8        1        0      active sync   /dev/sda1
           1       8       17        1      active sync   /dev/sdb1
    just curious, what do you get when you do a fdisk -l?

  9. #19
    Untangle Ninja sky-knight's Avatar
    Join Date
    Apr 2008
    Location
    Phoenix, AZ
    Posts
    26,230

    Default

    Code:
    Disk /dev/sda: 85.8 GB, 85899345920 bytes
    255 heads, 63 sectors/track, 10443 cylinders
    Units = cylinders of 16065 * 512 = 8225280 bytes
    Disk identifier: 0x0008bd47
    
       Device Boot      Start         End      Blocks   Id  System
    /dev/sda1   *           1        9726    78124063+  fd  Linux raid autodetect
    /dev/sda2            9727       10443     5759302+   5  Extended
    /dev/sda5            9727       10443     5759271   82  Linux swap / Solaris
    
    Disk /dev/sdb: 85.8 GB, 85899345920 bytes
    255 heads, 63 sectors/track, 10443 cylinders
    Units = cylinders of 16065 * 512 = 8225280 bytes
    Disk identifier: 0x000ada08
    
       Device Boot      Start         End      Blocks   Id  System
    /dev/sdb1   *           1        9726    78124063+  fd  Linux raid autodetect
    /dev/sdb2            9727       10443     5759302+   5  Extended
    /dev/sdb5            9727       10443     5759271   82  Linux swap / Solaris
    
    Disk /dev/md0: 79.9 GB, 79998943232 bytes
    2 heads, 4 sectors/track, 19530992 cylinders
    Units = cylinders of 8 * 512 = 4096 bytes
    Disk identifier: 0x00000000
    
    Disk /dev/md0 doesn't contain a valid partition table
    I love the final line, obviously if it didn't have a valid partition table it wouldn't have booted! lol
    Rob Sandling, BS:SWE, MCP
    NexgenAppliances.com
    Phone: 866-794-8879 x201
    Email: support@nexgenappliances.com

  10. #20
    Untangle Ninja sky-knight's Avatar
    Join Date
    Apr 2008
    Location
    Phoenix, AZ
    Posts
    26,230

    Default

    An interesting tidbit, I removed one of the virtual drives and it boots fine.

    Code:
                                                                                                         
    # /etc/fstab: static file system information.
    #
    # <file system> <mount point>   <type>  <options>       <dump>  <pass>
    proc            /proc           proc    defaults        0       0
    /dev/md0        /               ext3    errors=remount-ro 0       1
    /dev/sda5       none            swap    sw              0       0
    /dev/sdb5       none            swap    sw              0       0
    /dev/hdc        /media/cdrom0   udf,iso9660 user,noauto     0       0
    /dev/fd0        /media/floppy0  auto    rw,user,noauto  0       0
    
    Disk /dev/sda: 85.8 GB, 85899345920 bytes
    255 heads, 63 sectors/track, 10443 cylinders
    Units = cylinders of 16065 * 512 = 8225280 bytes
    Disk identifier: 0x0008bd47
    
       Device Boot      Start         End      Blocks   Id  System
    /dev/sda1   *           1        9726    78124063+  fd  Linux raid autodetect
    /dev/sda2            9727       10443     5759302+   5  Extended
    /dev/sda5            9727       10443     5759271   82  Linux swap / Solaris
    
    Disk /dev/md0: 79.9 GB, 79998943232 bytes
    2 heads, 4 sectors/track, 19530992 cylinders
    Units = cylinders of 8 * 512 = 4096 bytes
    Disk identifier: 0x00000000
    
    Disk /dev/md0 doesn't contain a valid partition table
    
    /dev/md0:
            Version : 00.90
      Creation Time : Tue Jun 14 15:19:50 2011
         Raid Level : raid1
         Array Size : 78123968 (74.50 GiB 80.00 GB)
      Used Dev Size : 78123968 (74.50 GiB 80.00 GB)
       Raid Devices : 2
      Total Devices : 1
    Preferred Minor : 0
        Persistence : Superblock is persistent
    
        Update Time : Tue Jun 14 18:57:37 2011
              State : clean, degraded
     Active Devices : 1
    Working Devices : 1
     Failed Devices : 0
      Spare Devices : 0
    
               UUID : 711aaf3d:b51cd6b0:56ff4bb0:38d8c331
             Events : 0.122
    
        Number   Major   Minor   RaidDevice State
           0       8        1        0      active sync   /dev/sda1
           1       0        0        1      removed
    I need to dig up a pair of 80gb SATA drives for a bare install here... so I can yank one out while it's running to see how the OS handles the loss of the swap partition. Dealing with that over a reboot seems reasonable, as udev likely just doesn't mount it. But, I can't imagine yanking 50% of the box's swap space mid stream is healthy for a running server.

    After I think about this... but my OpenFiler box runs like this and it's lost a drive and it didn't seem to care. I don't remember if I rebooted it or not.

    My partitioning is based on default Ubuntu and Debian, not Untangle.

    You really want your swap at the end, I think Untangle does what they do just because it's easier to script making an 80gb partition, a 2x ram partition, and a whatever is left partition. Of course, the tools allow you to make a partition at the "end" so that just seems silly to me. Supposedly you get better performance on the end of a drive.
    Last edited by sky-knight; 06-14-2011 at 07:24 PM.
    Rob Sandling, BS:SWE, MCP
    NexgenAppliances.com
    Phone: 866-794-8879 x201
    Email: support@nexgenappliances.com

Page 2 of 3 FirstFirst 123 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  

SEO by vBSEO 3.6.0 PL2