[DRBD-user] tuning

[Papp Tamas]

hi All,


Can I tune somehow this up?

global {
    usage-count yes;
}
common {
  syncer { rate 100M; }
}
resource data {
  protocol C;
  handlers {
    pri-on-incon-degr "echo o > /proc/sysrq-trigger ; halt -f";
    pri-lost-after-sb "echo o > /proc/sysrq-trigger ; halt -f";
    local-io-error "echo o > /proc/sysrq-trigger ; halt -f";
    outdate-peer "/usr/lib/heartbeat/drbd-peer-outdater -t 5";
    pri-lost "echo pri-lost. Have a look at the log files. | mail -s 
'DRBD Alert' root";
    split-brain "echo split-brain. drbdadm -- --discard-my-data connect 
$DRBD_RESOURCE ? | mail -s 'DRBD Alert' root";
    out-of-sync "echo out-of-sync. drbdadm down $DRBD_RESOURCE. drbdadm 
::::0 set-gi $DRBD_RESOURCE. drbdadm up $DRBD_RESOURCE. | mail -s 'DRBD 
Alert' root";
  }
  startup {
  }
  disk {
    on-io-error   detach;
  }
  net {
    max-buffers     2048;
    unplug-watermark   2048;
    cram-hmac-alg "sha1";
    shared-secret "asdjasdfasdfsd234kfj";
    after-sb-0pri disconnect;
    after-sb-1pri disconnect;
    after-sb-2pri disconnect;
    rr-conflict disconnect;
  }
  syncer {
    rate 100M;
    al-extents 257;
    cpu-mask 15;
  }
  on meta1 {
    device     /dev/drbd0;
    disk       /dev/md3;
    address    192.168.254.1:7788;
    flexible-meta-disk  internal;
  }
  on meta2 {
    device    /dev/drbd0;
    disk      /dev/md3;
    address   192.168.254.2:7788;
    meta-disk internal;
  }
}


Still the node 'meta2' is not connected, but already it's so slow, it 
cannot acting as a Lustre MDS.

CentOS 5.2 (upgraded from 5.0),
Linux meta1 2.6.18-prep #2 SMP Sat Sep 20 17:09:35 CEST 2008 x86_64 
x86_64 x86_64 GNU/Linux

GIT-hash: 3e69822d3bb4920a8c1bfdf7d647169eba7d2eb4 build by 
root at backup4.cube.intra, 2008-09-20 17:35:38
 0: cs:Connected st:Primary/Secondary ds:UpToDate/Diskless C r---
    ns:0 nr:0 dw:68433384 dr:267234439 al:57486 bm:8070 lo:0 pe:0 ua:0 
ap:0 oos:171866012


v8.2.6

With v0.7.25 it was much more faster.

What else could I say?



Thank you,

tamas