Hi all
This new thread is a spin-off of thread http://www.vmware.com/community/thread.jspa?threadID=63666 as I think that this could be a different issue.
I have a classroom with several pcs running VMware server on Mandriva 2007 32 bits that I use to teach. VMware is configured with a bridged vmnet0 to the only one eth0 interface, and Mandriva get its network configuration through DHCP.
Everything installed without problems, and my pupils were a couple of days loading VMs and playing happily with them.
But suddenly at the start of a class last Friday, everyone of them started to suffer the same symptoms: VMware server hangs on start, and it last 40-60 seconds to load, instead of the usual 4-6 seconds. And when it loads, the GUI is rather unresponsible and everything is "sluggish".
I discovered that if you shut down the network interface that the VM is bridged to, VMware server start snappy and everything seems to function OK... until you reconnect the bridged interface. An this happens in all the servers (7) that I have in the student's pcs.
I had the clear impression that the server hangs waiting "something" on the bridged vmnet0 and, as soon as the eth0 interface it is bridged to is shutted down in the host OS, the server continues to function as usual.
I have been checking /var/log/messages today and have found something that appears to be the culprit of the "hangins" of VMware server.
Next follows some of /var/log/messages output as I pressed the reset button on a running VM. VMware server hangs at 18:42:12
Lines marked with double asterisks and dashes are the hot spots.
\----
Dec 5 18:41:48 localhost kernel: device eth0 entered promiscuous mode
Dec 5 18:41:48 localhost kernel: audit(1165340508.156:16): dev=eth0 prom=256 old_prom=0 auid=4294967295
Dec 5 18:41:48 localhost kernel: device lo entered promiscuous mode
Dec 5 18:41:48 localhost kernel: audit(1165340508.164:17): dev=lo prom=256 old_prom=0 auid=4294967295
Dec 5 18:41:50 localhost kernel: device eth0 left promiscuous mode
Dec 5 18:41:50 localhost kernel: audit(1165340510.297:18): dev=eth0 prom=0 old_prom=256 auid=4294967295
Dec 5 18:41:50 localhost kernel: device lo left promiscuous mode
Dec 5 18:41:50 localhost kernel: audit(1165340510.317:19): dev=lo prom=0 old_prom=256 auid=4294967295
Dec 5 18:41:50 localhost kernel: device eth0 entered promiscuous mode
Dec 5 18:41:50 localhost kernel: audit(1165340510.369:20): dev=eth0 prom=256 old_prom=0 auid=4294967295
Dec 5 18:41:59 localhost kernel: bridge-eth0: enabled promiscuous mode
\**** Dec 5 18:42:12 localhost kernel: bridge-eth0: disabled promiscuous mode
\----- Dec 5 18:42:57 localhost kernel: device eth0 left promiscuous mode
Dec 5 18:42:57 localhost kernel: audit(1165340577.677:21): dev=eth0 prom=0 old_prom=256 auid=4294967295
\----
From 18:42:12 to 18:42:57 VMware server is "hanged". It appears to me that the transition from promiscuous to non-promiscuous mode in eth0 is causing VMware server to "hang on something".
I can reproduce those "hangins" simply by pressing the reset or stop buttons in the GUI for the running VM.
In the following output VMware server "hangs" at 18:48:37 and 18:50:04
\----
Dec 5 18:47:37 localhost kernel: bridge-eth0: disabled promiscuous mode
Dec 5 18:48:37 localhost kernel: device eth0 left promiscuous mode
\**** Dec 5 18:48:37 localhost kernel: audit(1165340917.118:33): dev=eth0 prom=0 old_prom=256 auid=4294967295
\----- Dec 5 18:50:03 localhost kernel: device eth0 entered promiscuous mode
Dec 5 18:50:03 localhost kernel: audit(1165341003.339:34): dev=eth0 prom=256 old_prom=0 auid=4294967295
Dec 5 18:50:03 localhost kernel: device lo entered promiscuous mode
Dec 5 18:50:03 localhost kernel: audit(1165341003.351:35): dev=lo prom=256 old_prom=0 auid=4294967295
Dec 5 18:50:04 localhost kernel: device eth0 left promiscuous mode
Dec 5 18:50:04 localhost kernel: audit(1165341004.463:36): dev=eth0 prom=0 old_prom=256 auid=4294967295
Dec 5 18:50:04 localhost kernel: device lo left promiscuous mode
Dec 5 18:50:04 localhost kernel: audit(1165341004.487:37): dev=lo prom=0 old_prom=256 auid=4294967295
Dec 5 18:50:04 localhost kernel: device eth0 entered promiscuous mode
\***** Dec 5 18:50:04 localhost kernel: audit(1165341004.531:38): dev=eth0 prom=256 old_prom=0 auid=4294967295
\----- Dec 5 18:51:22 localhost kernel: bridge-eth0: enabled promiscuous mode
Dec 5 18:51:43 localhost kernel: bridge-eth0: disabled promiscuous mode
Dec 5 18:51:43 localhost kernel: bridge-eth0: enabled promiscuous mode
Dec 5 18:51:43 localhost kernel: hdc: packet command error: status=0x51 \{ DriveReady SeekComplete Error }
Dec 5 18:51:43 localhost kernel: hdc: packet command error: error=0x54 \{ AbortedCommand LastFailedSense=0x05 }
Dec 5 18:51:43 localhost kernel: ide: failed opcode was: unknown
\----
As I said before, if I shut down eth0, VMware server starts up rather quickly (4-6 seconds). Also, if I start VMware with eth0 enabled, the server last between 40-60 seconds to show the GUI. If I disable eth0 while VMware server is starting up, the server appears to response with normal speed.
I have to note that I also run VMware server in the classroom firewall/server (with to NICs and the same version of Mandriva) without any of this problems.
Does anyone here have a clue on this scenario? What/where can I look to find clues on this?
Thank you all