Thursday, January 3, 2013

Now to nervously wait as this raid does a live rebuild

Added 2 drives to my raid 5 array.  going from 6 1TB drives to 8 1TB drives.    Rebuilding now.    This many drives in raid 5 make me a little nervous.  My kernel is too old to grow to raid 6.  I'm sure I'll be fine, but still nerve wracking, especially with my luck.


[root@borg temp]# mdadm --detail /dev/md0
/dev/md0:
        Version : 0.90
  Creation Time : Sat Sep 20 02:15:28 2008
     Raid Level : raid5
     Array Size : 4883799680 (4657.55 GiB 5001.01 GB)
  Used Dev Size : 976759936 (931.51 GiB 1000.20 GB)
   Raid Devices : 6
  Total Devices : 8
Preferred Minor : 0
    Persistence : Superblock is persistent

    Update Time : Thu Jan  3 21:13:05 2013
          State : clean
 Active Devices : 6
Working Devices : 8
 Failed Devices : 0
  Spare Devices : 2

         Layout : left-symmetric
     Chunk Size : 64K

           UUID : 11f961e7:0e37ba39:2c8a1552:76dd72ee
         Events : 0.1440052

    Number   Major   Minor   RaidDevice State
       0       8       96        0      active sync   /dev/sdg
       1       8       16        1      active sync   /dev/sdb
       2       8       32        2      active sync   /dev/sdc
       3       8       48        3      active sync   /dev/sdd
       4       8      112        4      active sync   /dev/sdh
       5       8       80        5      active sync   /dev/sdf

       6       8      128        -      spare   /dev/sdi
       7       8      160        -      spare   /dev/sdk
[root@borg temp]# 
[root@borg temp]# 
[root@borg temp]# df -hl
Filesystem            Size  Used Avail Use% Mounted on
/dev/sda3             433G   12G  400G   3% /
/dev/sda1             190M   25M  156M  14% /boot
tmpfs                 3.8G   48K  3.8G   1% /dev/shm
/dev/md0              4.5T  3.3T  995G  78% /raid1
[root@borg temp]# 
[root@borg temp]# 
[root@borg temp]# 
[root@borg temp]# cat simplebenchmark.sh 
echo
dd if=/dev/zero of=test.bin bs=5000000 count=1000
echo
echo
dd of=/dev/null if=test.bin bs=5000000
echo

[root@borg temp]# ./simplebenchmark.sh 

1000+0 records in
1000+0 records out
5000000000 bytes (5.0 GB) copied, 24.2193 s, 206 MB/s


1000+0 records in
1000+0 records out
5000000000 bytes (5.0 GB) copied, 3.03028 s, 1.7 GB/s

[root@borg temp]# 
[root@borg temp]# 
[root@borg temp]# 
[root@borg temp]# 
[root@borg temp]# mdadm --grow /dev/md0 --raid-devices=8
mdadm: Need to backup 1344K of critical section..
mdadm: ... critical section passed.
[root@borg temp]# 
[root@borg temp]# 
[root@borg temp]# 
[root@borg temp]# 
[root@borg temp]# 
[root@borg temp]# mdadm --detail /dev/md0
/dev/md0:
        Version : 0.91
  Creation Time : Sat Sep 20 02:15:28 2008
     Raid Level : raid5
     Array Size : 4883799680 (4657.55 GiB 5001.01 GB)
  Used Dev Size : 976759936 (931.51 GiB 1000.20 GB)
   Raid Devices : 8
  Total Devices : 8
Preferred Minor : 0
    Persistence : Superblock is persistent

    Update Time : Thu Jan  3 21:16:08 2013
          State : clean, recovering
 Active Devices : 8
Working Devices : 8
 Failed Devices : 0
  Spare Devices : 0

         Layout : left-symmetric
     Chunk Size : 64K

 Reshape Status : 0% complete
  Delta Devices : 2, (6->8)

           UUID : 11f961e7:0e37ba39:2c8a1552:76dd72ee
         Events : 0.1440210

    Number   Major   Minor   RaidDevice State
       0       8       96        0      active sync   /dev/sdg
       1       8       16        1      active sync   /dev/sdb
       2       8       32        2      active sync   /dev/sdc
       3       8       48        3      active sync   /dev/sdd
       4       8      112        4      active sync   /dev/sdh
       5       8       80        5      active sync   /dev/sdf
       6       8      128        6      active sync   /dev/sdi
       7       8      160        7      active sync   /dev/sdk
[root@borg temp]# 


Will be interesting to see the IO speed after this.  I wonder if I've already reached a plateau, or if I can actually get more out of it.  I'm mostly doing this for disk space and not performance mind you, but always nice to get some extra performance too. 

There's also a hot spare that's not added but when I was doing testing it had lot of smart errors so think I'll have to RMA it.    Using all WD Blacks 1TB. I missed out on the cheap 3TB drives pre flood, but at slightly under 100 bucks these 1TB were hard to resist. NCIX had a sale.  All my slots are full now in my server.    This will give me about 1.7TB of extra space. By the time I remotely come close to using all that I'll probably have built a SAN anyway.  I rarely delete data, I just add more disk space. :D

No comments: