I don't understand this, but I have 2 suggestions/requests.  First, I would suggest you run strace on the mount to see where it is hanging.   Second, I would suggest you use CVIP in a different way:
   a) you are using the physical address of node 1 as your cvip;  it is recommended that you actually have a virtual address (i.e. not the physical address);  that way
       you have the chance of failing it over to another node if the first node fails.
   b) your gateway (in the cvip.conf) is the external physical address of node 1;   For NAT, i believe the gateway needs to be the physical address of the internal
       network (so node 2, etc. can go thru node 1 to get to the outside world).
 
I strongly suggest that something in the way you configured your CVIP is causing the issue but I don't exactly understand what/how.
 
 
bruce
 
-----Original Message-----
From: ssic-linux-devel-admin@lists.sourceforge.net [mailto:ssic-linux-devel-admin@lists.sourceforge.net] On Behalf Of Smith, Stan
Sent: Wednesday, December 29, 2004 3:58 PM
To: ssic-linux-devel@lists.sourceforge.net
Subject: [SSI-devel] R1.2.0-FC2: ha-lvs slows NFS mount/umounts to a snail's pace

Summary:
========
  R1.2.0-FC2: When ha-lvs is enabled, '/etc/cvip.conf' is present, an NFS mount from the init node (the only node up) take approx. 5 minutes. When ha-lvs is disabled (system rebooted - no '/etc/cvip.conf'), same NFS mount occurs in seconds?

Details:
======
Vanilla R1.2.0 FC2 'fresh' install.

The init node is the only being used, all other nodes are powered-down.

Init node uses eth0 as the gateway to the corporate net.

eth1 is the 'private' ics network.

Init node NFS mounts 'yoda:/usr/local/src/IBA' from /etc/fstab; takes about
5 seconds for the boot mount script to complete the NFS mount successfully.

All works as expected until ha-lvs comes into the picture.
When an '/etc/cvip.conf' file is present (see below) the NFS mount takes
about 5 minutes instead of the 5 seconds when the /etc/cvip.conf file is NOT
present. The good news is the mount eventually occurs, bad news is it takes
5 minutes.

If the NFS mount is removed from /etc/fstab, init node boots normally with ha-lvs enabled. If one
attempts to NFS mount after the init node is up, the mount takes an
abnormally long time (again 5 minutes); same as if the NFS mount was in
/etc/fstab.

Unmounting an NFS mount also takes an abnormally long time if ha-lvs is
enabled; works fine if ha-lvs is NOT enabled.

Suggestions on how to remedy and/or further debug the situation?

thanks,

stan.


/etc/fstab NFS entry:
yoda:/usr/local/src/IBA /usr/local/src/IBA nfs rsize=8192,wsize=8192,intr,timeo=16,node=1


mount
/dev/1/sda2 on / type ext3 (rw)
none on /proc type proc (rw)
none on /sys type sysfs (rw)
none on /dev/pts type devpts (rw,gid=5,mode=620)
devfs on /dev type devfs (rw)
/dev/1/sda1 on /boot type ext3 (rw)
/dev/1/sdb1 on /home type ext3 (rw)
yoda:/usr/local/src/IBA on /usr/local/src/IBA type nfs (rw,rsize=8192,wsize=8192,intr,timeo=16,addr=9.7.163.85)


ip route
10.10.3.0/24 dev eth1  proto kernel  scope link  src 10.10.3.20
9.7.163.0/24 dev eth0  proto kernel  scope link  src 9.7.163.127
default via 9.7.163.251 dev eth0

ifconfig eth0
eth0      Link encap:Ethernet  HWaddr 00:D0:B7:B6:D6:02 
          inet addr:9.7.163.127  Bcast:9.7.163.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:5516 errors:0 dropped:0 overruns:0 frame:0
          TX packets:749 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:4065857 (3.8 Mb)  TX bytes:83459 (81.5 Kb)
          Interrupt:18 Base address:0x5400 Memory:e8101000-e8101038

cat /etc/sysconfig/network-scripts/ifcfg-eth0
# Intel Corp.|82557/8/9 [Ethernet Pro 100]
DEVICE=eth0
HWADDR=00:D0:B7:B6:D6:02
BOOTPROTO=dhcp
ONBOOT=yes
TYPE=Ethernet
DHCP_HOSTNAME=venus

ifconfig eth1
eth1      Link encap:Ethernet  HWaddr 00:03:47:3F:CA:33 
          inet addr:10.10.3.20  Bcast:10.10.3.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:4 errors:0 dropped:0 overruns:0 carrier:4
          collisions:0 txqueuelen:1000
          RX bytes:0 (0.0 b)  TX bytes:256 (256.0 b)
          Interrupt:21 Memory:ea020000-ea040000

cat /etc/sysconfig/network-scripts/ifcfg-eth1
# Intel Corp.|82543GC Gigabit Ethernet Controller (Copper)
DEVICE=eth1
BOOTPROTO=static
ONBOOT=yes
TYPE=Ethernet
HWADDR=00:03:47:3F:CA:33
IPADDR=10.10.3.20
NETMASK=255.255.255.0


cat /etc/cvip.conf

<?xml version="1.0"?>
<cvips>
        <routing>NAT</routing>
        <cvip>
                <ip_addr>9.7.163.127</ip_addr>
                <gateway>9.7.163.127</gateway>
                <director_node>
                        <node_num>1</node_num>
                        <garp_interface>eth0</garp_interface>
                        <sync_interface>eth1</sync_interface>
                </director_node>
                <real_server_node>
                        <node_num>2</node_num>
                </real_server_node>
                <real_server_node>
                        <node_num>3</node_num>
                </real_server_node>
                <real_server_node>
                        <node_num>4</node_num>
                </real_server_node>
                <real_server_node>
                        <node_num>5</node_num>
                </real_server_node>
                <real_server_node>
                        <node_num>6</node_num>
                </real_server_node>
                <real_server_node>
                        <node_num>7</node_num>
                </real_server_node>
                <real_server_node>
                        <node_num>8</node_num>
                </real_server_node>
                <real_server_node>
                        <node_num>9</node_num>
                </real_server_node>
                <real_server_node>
                        <node_num>10</node_num>
                </real_server_node>
                <real_server_node>
                        <node_num>11</node_num>
                </real_server_node>
                <real_server_node>
                        <node_num>12</node_num>
                </real_server_node>
                <real_server_node>
                        <node_num>13</node_num>
                </real_server_node>
                <real_server_node>
                        <node_num>14</node_num>
                </real_server_node>
                <real_server_node>
                        <node_num>15</node_num>
                </real_server_node>
                <real_server_node>
                        <node_num>16</node_num>
                </real_server_node>
                <real_server_node>
                        <node_num>17</node_num>
                </real_server_node>
                <real_server_node>
                        <node_num>18</node_num>
                </real_server_node>
        </cvip>
</cvips>