[lvs-users] LVS-DR and scp

Scooter Morris scooter at cgl.ucsf.edu
Tue Dec 1 23:02:47 GMT 2009


Hi all,
     I have a two-node LVS configuration that I'm using as a front-end 
to a three-node cluster.  It works really well for ssh, http, https, 
etc.   But, when a user tries to use scp to copy a large file or 
directory to the cluster, it commonly stalls and doesn't complete.  scp 
directly to any node of the cluster works fine, either from outside the 
network or from inside it.  Things only crawl when using scp to go 
through the redirectors.  Here are the relevant parts of the 
configuration file:

service = lvs
primary = xxx.xxx.xxx.30
backup = xxx.xxx.xxx.2
backup_active = 1
heartbeat = 1
heartbeat_port = 539
keepalive = 12
deadtime = 4
debug_level = NONE
network = direct
virtual plato-ssh {
      active = 1
      address = xxx.xxx.xxx.29 eth0:1
      vip_nmask = 255.255.255.0
      port = 22
      expect = "SSH-2.0"
      send = "SSH-2.0-Probe"
      use_regex = 0
      load_monitor = rup
      scheduler = wlc
      protocol = tcp
      timeout = 60
      reentry = 60
      quiesce_server = 0
      persistent = 600
      server node1 {
          address = xxx.xxx.xxx.36
          active = 1
          weight = 20
      }
      server node2 {
          address = xxx.xxx.xxx.34
          active = 1
          weight = 20
      }
      server node3 {
          address = xxx.xxx.xxx.32
          active = 1
          weight = 1
      }
}

This is with kernel version 2.6.18-164.2.1.el5 and I'm running piranha 
version 0.8.4-11.el5.  Has anyone seen anything like this before or have 
any ideas what I can do to try to debug it?

Thanks!

-- scooter

-------------- next part --------------
A non-text attachment was scrubbed...
Name: scooter.vcf
Type: text/x-vcard
Size: 378 bytes
Desc: not available
Url : http://lists.graemef.net/pipermail/lvs-users/attachments/20091201/c38703ca/attachment.vcf 


More information about the lvs-users mailing list