Bridged adapter fails with DHCP

[caramba]

Tests performed using:
- Version: 4.3.6
- Guest VM: CentOS 6.5 x64 Minimal installation
- Host OSs (with latest patches)
- Linux: CentOS 6.5 x64
- Windows: 7 x64

Test case 1:
1. created a new Guest VM with only Bridged adapter on Linux host and booted successfully with only lo adapter
2. Changed BOOT=no to BOOT=yes in /etc/sysconfig/network-scripts/ifcfg-eth0
3. Rebooted and VM rebooted successfully with lo and eth0 adapters. ifconfig identified a valid DHCP provided IP address for eth0
4. powered the VM down via the poweroff command
5. Snapshotted the VM (at this point everything was working)
6. Tried to start the VM again on the Linux host some hours later ...and it sticks on "Determining IP information for eth0" ...for like ever.
7. Eventually it times out? and completes the boot. eth0 shows as active but no IP is reflected. (I tried a million things to get it to reflect but nothing worked)
8. Restored the original snapshot from 5 above and copied the whole folder tree across to Windows 7 host.
9. Started the VM there ....same problem as explained in steps 6 & 7.

So I thought well, what is common above? A. the guest VM; and B. the DHCP server configuration.

Test case 2:
1. created a new Guest VM with only Bridged adapter on Windows 7 host and booted successfully with only lo adapter
2. Changed BOOT=no to BOOT=yes in /etc/sysconfig/network-scripts/ifcfg-eth0
3. Rebooted and VM rebooted successfully with lo and eth0 adapters. ifconfig identified a valid DHCP provided IP address for eth0
4. powered the VM down via the poweroff command
5. Snapshotted the VM (at this point everything was working)
6. Copied the VM folder tree across to the Linux host and booted successfully. ifconfig identified a valid DHCP provided IP address for eth0
7. Powered the VM down.
8. Tried to start the VM again on the Linux host some hours later ...and it sticks on "Determining IP information for eth0" ...for like ever.
9. Took the Windows 7 host (a laptop) home and attached to an entirely different network with a different DHCP server.
10 Started the VM snapshotted in step 5 ...and it sticks on "Determining IP information for eth0" ...for like ever.

So in summary:
A CentOS 6.5 x64 VM with a bridged adapter configured for DHCP appears to work initially but then for some reason fails after a period of time - kind of like it isn't refreshed and times out.
This problem is consistent for the above configuration built on 2 different host operating systems and tested on 2 different networks with 2 different DHCP servers.

Additional note:
It may also be worth mentioning that changing from DHCP to static IP resolves the problem and the bridged network then works fine in both the above cases

I can therefore only conclude that there is a problem with the bridged adapter DHCP implementation in 4.3.6

I do hope this is useful

[noteirak]

Thank you for the very detailed info.

The best way to get this info the devs is to open a ticket (or add to an already existing one about this bug) on the BugTracker.
Make sure you include this content and the VM log file when you reproduce the issue.

Moved this post to Linux guests.