Bug 10165 - ping_pong - network extremely slow
ping_pong - network extremely slow
Status: RESOLVED INVALID
Product: CTDB 2.5.x or older
Classification: Unclassified
Component: ctdb
unspecified
All Linux
: P5 normal
: ---
Assigned To: Michael Adam
Samba QA Contact
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2013-09-25 10:00 UTC by edgar.wiest
Modified: 2013-09-25 15:19 UTC (History)
0 users

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description edgar.wiest 2013-09-25 10:00:59 UTC
hello

when I run ping_pong on a node, then the whole network is slow or falls out completely.

what ist wrong?

ping_pong -rw /clusterfs/data/test.odt 3

system: centos 6.4
cluster file system: gfs2

ctdb-1.0.114.5-3.el6.x86_64
gfs2-utils-3.0.12.1-49.el6_4.2.x86_64

2 nodes

/etc/cluster.conf on all nodes

<cluster name="test" config_version="3">
<cman two_node="1" expected_votes="1"/>
<clusternodes>
<clusternode name="cluster1.xxx.local" votes="1" nodeid="1">
        <fence>
                <method name="single">
                        <device name="manual" ipaddr="172.18.10.45"/>
                </method>
        </fence>
</clusternode>
<clusternode name="cluster2.xxx.local" votes="1" nodeid="2">
        <fence>
                <method name="single">
                        <device name="manual" ipaddr="172.18.10.47"/>
                </method>
        </fence>
</clusternode>
</clusternodes>
<fence_daemon clean_start="1" post_fail_delay="0" post_join_delay="3"/>
<fencedevices>
        <fencedevice name="manual" agent="fence_manual"/>
</fencedevices>
</cluster>

ctdb config:
CTDB_RECOVERY_LOCK="/clusterfs/ctdb.lock"
CTDB_PUBLIC_INTERFACE=eth0
CTDB_PUBLIC_ADDRESSES=/etc/ctdb/public_addresses
CTDB_MANAGES_SAMBA=yes
CTDB_SAMBA_CHECK_PORTS="445"
CTDB_MANAGES_WINBIND=yes
CTDB_MANAGES_VSFTPD=no
CTDB_MANAGES_ISCSI=no
CTDB_MANAGES_NFS=no
CTDB_MANAGES_HTTPP=no
CTDB_NOTIFY_SCRIPT=/etc/ctdb/notify.sh
CTDB_DEBUGLEVEL=3

no erros in log.ctdb

You need more information?
Comment 1 Volker Lendecke 2013-09-25 15:19:18 UTC
Assuming /clusterfs/data/test.odt is on a gfs mount, this is nothing Samba or ctdb could do anything about. Please contact the GFS people. If /clusterfs/data/test.odt is not a direct GFS mount but on a cifs client mounting a clusterfs remotely via Samba and ctdb, please re-open this bug.

Thanks,

Volker