[DRBD-user] Expected sync speed for a 425gb volume?

Nate Carlson natecars at real-time.com
Wed Apr 6 18:37:18 CEST 2005

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


Hey all,

I'm setting up DRBD between two identical boxes (Xeon 2.8ghz, 3ware 
8506-4LP's, 4 160gb SATA drives in RAID5), with a pair of GigE cards 
bonded together (iptraf says I can get about 1.8gbits/sec across the 
link). I'm running Debian Testing with a 2.6.10 kernel and drbd 0.7.10.

I've set up a ~425gb volume, with the following settings (snipped 
startup and disk):

resource data {
  protocol C;
  net {
    max-buffers     4096;
    max-epoch-size  1024;
  }
  syncer {
    rate 600M;
    group 1;
    al-extents 257;
  }
  on server1 {
    device     /dev/drbd0;
    disk       /dev/sda5;
    address    10.50.3.1:7788;
    meta-disk  internal;
  }
  on server2 {
    device    /dev/drbd0;
    disk      /dev/sda5;
    address   10.50.3.2:7788;
    meta-disk internal;
  }
}

I can start the disk ok, but I've been running mkfs.ext3 on it for 1h45m 
now, and it's still not done:

Writing inode tables: 2282/3434

Also, the sync is going very slow:

 1: cs:SyncSource st:Primary/Secondary ld:Consistent
    ns:22596892 nr:0 dw:4267388 dr:18345948 al:1947 bm:1133 lo:2 pe:74 ua:4096 ap:47
        [=>..................] sync'ed:  5.2% (402553/424334)M
        finish: 13:28:15 speed: 8,448 (6,840) K/sec

Both boxes are also showing a high load average (server1 is primary, and
is at ~2; server2 is secondary, and is at ~1).

Is this to be expected, or am I missing some important performance tweaks?

Thanks!

-- 
Nate Carlson <natecars at real-time.com>   | Phone : (952)943-8700
http://www.real-time.com                | Fax   : (952)943-8500




More information about the drbd-user mailing list