Hello all!
Today, I do some "bonnie++" tests on our NFS cluster (active/passive
with ha 2.1.2 and drbd 0.8.4).
I used bonnie++ on a client for generating much more traffic and load.
client: #mount -t nfs -o udp nfsvip:/export/uploads /mnt
client: #bonnie -d /mnt/ -s 750 -v 8 -p 2
With this test I got a load between 5 and 12. The nfs share grows up
to 5 GB.
For produce the failover I removed with rmmod the nfsd kernel mod.
(~: #rmmod -f nfsd).
The result is a very nice crash:)
Heartbeat switched as expected but the drbd failover crashed under the
high load.
I need a complete reconfiguration of drbd with a new sync.
Also the client crashed, I must reset the client via iLO.
The cluster was unusable;(
Without bonnie++ all failover scenarios works fine.
(reconnect client via udp/tcp, failover, etc)
Have someone experience with high load/traffic network storage
solutions?
IMHO is a active/active environment a better solution.
Should I use NFS or is GFS/OCFS2 the better way?
Or is my configuration "bullshit"?
~: #cat /etc/drbd.conf
- - - - - - - - - - - -
global { usage-count no; } # sorry, secure backend, no www;(
common { syncer { rate 100M; } }
resource drbd0
{
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/sbin/drbd-peer-outdater";
}
startup
{
degr-wfc-timeout 20;
}
syncer
{
rate 100M;
}
net
{
cram-hmac-alg sha1;
shared-secret "FooFunFactory";
}
on nfs00001 {
device /dev/drbd0;
disk /dev/cciss/c0d0p5;
address 1.1.1.1:7788;
meta-disk internal;
}
on nfs00002 {
device /dev/drbd0;
disk /dev/cciss/c0d0p5;
address 1.1.1.2:7788;
meta-disk internal;
}
}
- - - - - - - - - - - - - - - -
About feedback I would be very glad!
Regards
Andre
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
