VirtualBox

Ticket #4950 (closed defect: fixed)

Opened 5 years ago

Last modified 5 years ago

VirtualBox v3.0.6 lock up with high network traffic => fixed in SVN

Reported by: half12 Owned by:
Priority: major Component: network
Version: VirtualBox 3.0.6 Keywords: network system lock up
Cc: 7willlows@… Guest type: Solaris
Host type: Solaris

Description

Solaris Host : Dual Opteron 2218, 16GB RAM, 850GB sata disk, Supermicro H8DAE-2 motherboard

Installed VirtualBox 3.0.6 on my dual opteron system running Solaris 10 u5 and run a variety of Vbox, WinXP, SXCE b121 and not had any crashes or lock ups after nearly two hours. The updated AdditionsCD works correctly with SXCE b121 so that mouse interaction is seamless.

Initially my SXCE b121 VBox was configured with 2 cpus but I noticed that this was significantly slower than running under v2.2.4 with a single cpu. Reduced the number for cpus from 2 to 1 and restarted SXCE b121 and found that performance improved significantly. Have been running WinXP and SXCE b121 for around 5 hours, now started RedHat 4 the system is stable. Each of the VBox is configured to use Bridging. Please note that I have USB disabled in all VBox. I have been playing tv and radio content from the BBC as well as this video  http://www.youtube.com/watch?v=uGjiKrMO-7I So far, everything is working well.

Reproduced system lock up!! When trying to transfer around 10GB of files from Host to VBox client (both CentOS 5 and SXCE b121) the transfer rate using scp became slower and slower. Killed CentOS 5 transfer and tried with SXCE b121 this reached around 90% and then all VBox locked up but system still responsive, could switch between workspaces then eventually it froze and everything stopped.

Attachments

VBox.log.winxp Download (54.2 KB) - added by half12 5 years ago.
log of WinXP VBox
VBox.log.sxce-b121 Download (56.2 KB) - added by half12 5 years ago.
VBox.log.redhat4 Download (38.3 KB) - added by half12 5 years ago.
VBox.log.centos5 Download (39.8 KB) - added by half12 5 years ago.
vmstat.txt Download (7.2 KB) - added by half12 5 years ago.
vmstat 1 output
prstat Download (1.6 KB) - added by half12 5 years ago.
prstat
lockstat Download (15.3 KB) - added by half12 5 years ago.
lockstat

Change History

comment:1 Changed 5 years ago by Hachiman

Could you please attach log file.

Changed 5 years ago by half12

log of WinXP VBox

Changed 5 years ago by half12

Changed 5 years ago by half12

Changed 5 years ago by half12

comment:2 Changed 5 years ago by half12

Added logs files

comment:3 Changed 5 years ago by ramshankar

Could you please this latest test build:  http://www.virtualbox.org/download/testcase/VirtualBox-test-SunOS-3.0.7-r52830.tar.gz

(Note: This link will expire in 14 days from now)

comment:4 Changed 5 years ago by half12

Hi, Downloaded test v3.0.7 and installed at 13:30 and started testing.

1) Noticed significant impact on Solaris Host GUI response after installation!

2) Started WinXP VBox, attempted to run Stellarium 0.10.2 which failed to display to screen. After checking ~/.VirtualBox/Machines/winxp-ts3/LogsVBox.log found that it was an unsupported client. Attempted installation of GuestAdditions worked perfectly, no file permission weirdness (fixed Bug 5022). Note: Windows GUI in VBox is more responsive than host Solaris 10 GUI!

3) Started streaming of BBC radio on Windows XP, audio played perfectly drop outs did occur but that was due to internet congestion while Solaris 10 host downloaded files.

4) Started OpenSolaris VBox while Windows VBox running and started scp from VirtualBox Host 30GB of files as of 18:45 successfully transferred over 10GB without problem so far.

Note: No testing of the multi-processor functionality, only tested existing VBox.

comment:5 follow-up: ↓ 6 Changed 5 years ago by ramshankar

By "impact" I take it you mean the GUI has somehow become slower?

Yes, please complete single vCPU testing before trying multiple vCPUs. It would help in isolating problems easier. Thank you for testing.

comment:6 in reply to: ↑ 5 Changed 5 years ago by half12

Replying to ramshankar:

By "impact" I take it you mean the GUI has somehow become slower?

Yes, please complete single vCPU testing before trying multiple vCPUs. It would help in isolating problems easier. Thank you for testing.

Yes, mouse tracking is slower and jerky, pressing button for selection can result in it not being detected. Attempting to run OpenOffice while two VBox running took around 2.25 minutes to start (was listening to Elvis track on radio), also got Application not responding error message and forced it to quit.

comment:7 Changed 5 years ago by ramshankar

Could you please give me "vmstat 1" output for about 30 seconds while running the VMs and prstat output? And also output of "lockstat -kIW sleep 5" output. How long have the tests been running?

Changed 5 years ago by half12

vmstat 1 output

Changed 5 years ago by half12

prstat

Changed 5 years ago by half12

lockstat

comment:10 follow-up: ↓ 11 Changed 5 years ago by ramshankar

Thank you, how many cores does the system have (i.e. you can check psrinfo)

comment:11 in reply to: ↑ 10 Changed 5 years ago by half12

Replying to ramshankar:

Thank you, how many cores does the system have (i.e. you can check psrinfo)

Dual CPU, 2 cores per CPU (Dual Opteron 2218)

Stopped VBox to Host scp after 15GB. System stable only complaint is poor host response.

comment:12 Changed 5 years ago by ramshankar

Great! Thanks a lot for testing!

comment:13 Changed 5 years ago by half12

VBox multi-processor testing

1) VBox are slower to load

2) Tested two VBox both assigned 2 Processors and Vbox with single processor

SXCE b115 VBox, played Flash video within Firefox, audio experienced periodic buzz/blip sound. Ubuntu 9.04 worked initially, attempted to download Flash from www.adobe.com and mouse interaction became broken - unable to select anything, attempting to select with button click results in Wastebasket opening! Afterstarting OpenOffice and then the single processor VBox found that I was not able to get SXCE b115 VBox to respond to mouse clicks.

3) Find overall performance better of Host and VBox(s) when not using multiple processors within VBox.

comment:14 Changed 5 years ago by half12

Caused a system lockup while attempting to move a VBox Window during the boot up process (BIOS screen displayed), the Solaris 10 host was downloading the SXCE b123 at the time.

comment:15 follow-up: ↓ 16 Changed 5 years ago by ramshankar

This must have been a system reboot not lockup.

comment:16 in reply to: ↑ 15 Changed 5 years ago by half12

Replying to ramshankar:

This must have been a system reboot not lockup.

When I have experienced unexpected reboots the past (worked with SunOS since 1986) the time between lockup and reboot has been much quicker. As an aside, while performing a Solaris 10 host download with v3.0.7 installed it appeared to effect on the Firefox v3.5.2 download, while the download rate reported by Firefox was at the connection maximum of 235Kb/s the file size did not increase at the correct rate. However, after downgrading to v2.2.4 and attempting the download again; the 235Kb/s download rate did match the increase in file size.

comment:17 follow-up: ↓ 18 Changed 5 years ago by ramshankar

The time of reboot varies but it happens within a while, it's critical that this is a reboot and not the old hang while VMs were already running the system just hard froze all of a sudden.

Btw -  http://www.virtualbox.org/wiki/Core_dump <-- has instructions on how to get a core. In case of your reboot, if you had cores enabled I'm sure you'd have the core in /var/crash/<host>/ and the stack trace in "/var/adm/messages" file.

Thanks again for the testing!

comment:18 in reply to: ↑ 17 ; follow-up: ↓ 23 Changed 5 years ago by half12

Replying to ramshankar:

The time of reboot varies but it happens within a while, it's critical that this is a reboot and not the old hang while VMs were already running the system just hard froze all of a sudden.

Btw -  http://www.virtualbox.org/wiki/Core_dump <-- has instructions on how to get a core. In case of your reboot, if you had cores enabled I'm sure you'd have the core in /var/crash/<host>/ and the stack trace in "/var/adm/messages" file.

Thanks again for the testing!

Core Dumps were enabled

Messages error Sep 26 15:51:48 ts savecore: [ID 570001 auth.error] reboot after panic: BAD TRAP: type=e (#pf Page fault) rp=fffffe800085b250 addr=50 occurred in module "genunix" due to a NULL pointer dereference Sep 26 15:51:48 ts savecore: [ID 748169 auth.error] saving system crash dump in /var/crash/ts/*.0

comment:19 follow-up: ↓ 20 Changed 5 years ago by ramshankar

A full backtrace would be helpful, could you please look for the lines following this message you pasted on "/var/adm/messages". I just want to make sure it's the same as the one in  http://forums.virtualbox.org/viewtopic.php?f=11&t=20015&start=75#p100133

comment:20 in reply to: ↑ 19 Changed 5 years ago by half12

Replying to ramshankar:

A full backtrace would be helpful, could you please look for the lines following this message you pasted on "/var/adm/messages". I just want to make sure it's the same as the one in  http://forums.virtualbox.org/viewtopic.php?f=11&t=20015&start=75#p100133

Sorry, here is the full listing

Sep 26 15:48:06 ts unix: [ID 836849 kern.notice]
Sep 26 15:48:06 ts ^Mpanic[cpu1]/thread=ffffffff9a4ea520:
Sep 26 15:48:06 ts genunix: [ID 335743 kern.notice] BAD TRAP: type=e (#pf Page fault) rp=fffffe800085b250 addr=50 occurred in module "genunix" due to a NULL pointer dereference
Sep 26 15:48:06 ts unix: [ID 100000 kern.notice]
Sep 26 15:48:06 ts unix: [ID 839527 kern.notice] VirtualBox:
Sep 26 15:48:06 ts unix: [ID 753105 kern.notice] #pf Page fault
Sep 26 15:48:06 ts unix: [ID 532287 kern.notice] Bad kernel fault at addr=0x50
Sep 26 15:48:06 ts unix: [ID 243837 kern.notice] pid=1355, pc=0xfffffffffb9e65d5, sp=0xfffffe800085b340, eflags=0x10213
Sep 26 15:48:06 ts unix: [ID 211416 kern.notice] cr0: 8005003b<pg,wp,ne,et,ts,mp,pe> cr4: 6f0<xmme,fxsr,pge,mce,pae,pse>
Sep 26 15:48:06 ts unix: [ID 354241 kern.notice] cr2: 50 cr3: 428f42000 cr8: c
Sep 26 15:48:06 ts unix: [ID 592667 kern.notice]        rdi:                0 rsi:               31 rdx:                0
Sep 26 15:48:06 ts unix: [ID 592667 kern.notice]        rcx:                1  r8:                0  r9:                0
Sep 26 15:48:06 ts unix: [ID 592667 kern.notice]        rax:                0 rbx:             532f rbp: fffffe800085b370
Sep 26 15:48:06 ts unix: [ID 592667 kern.notice]        r10: fffffe800085b720 r11: ffffffff80b7adc8 r12:                0
Sep 26 15:48:06 ts unix: [ID 592667 kern.notice]        r13:               31 r14:                0 r15:                1
Sep 26 15:48:06 ts unix: [ID 592667 kern.notice]        fsb: fffffd7ffef10200 gsb: ffffffff99ddf800  ds:                0
Sep 26 15:48:06 ts unix: [ID 592667 kern.notice]         es:                0  fs:              1bb  gs:                0
Sep 26 15:48:06 ts unix: [ID 592667 kern.notice]        trp:                e err:                0 rip: fffffffffb9e65d5
Sep 26 15:48:06 ts unix: [ID 592667 kern.notice]         cs:               28 rfl:            10213 rsp: fffffe800085b340
Sep 26 15:48:06 ts unix: [ID 266532 kern.notice]         ss:               30
Sep 26 15:48:06 ts unix: [ID 100000 kern.notice]
Sep 26 15:48:06 ts genunix: [ID 655072 kern.notice] fffffe800085b160 unix:die+da ()
Sep 26 15:48:06 ts genunix: [ID 655072 kern.notice] fffffe800085b240 unix:trap+5e6 ()
Sep 26 15:48:06 ts genunix: [ID 655072 kern.notice] fffffe800085b250 unix:_cmntrap+140 ()
Sep 26 15:48:06 ts genunix: [ID 655072 kern.notice] fffffe800085b370 genunix:priv_policy+35 ()
Sep 26 15:48:06 ts genunix: [ID 655072 kern.notice] fffffe800085b380 genunix:secpolicy_net_config+34 ()
Sep 26 15:48:06 ts genunix: [ID 655072 kern.notice] fffffe800085b6a0 genunix:strioctl+1a45 ()
Sep 26 15:48:06 ts genunix: [ID 655072 kern.notice] fffffe800085b8f0 vboxflt:vboxNetFltSolarisAttachIp4+3c9 ()
Sep 26 15:48:06 ts genunix: [ID 655072 kern.notice] fffffe800085ba70 vboxflt:vboxNetFltOsInitInstance+5d8 ()
Sep 26 15:48:06 ts genunix: [ID 655072 kern.notice] fffffe800085baf0 vboxflt:vboxNetFltFactoryCreateAndConnect+227 ()
Sep 26 15:48:06 ts genunix: [ID 802836 kern.notice] fffffe800085bb70 fffffffff05b20a7 ()
Sep 26 15:48:06 ts genunix: [ID 802836 kern.notice] fffffe800085bba0 fffffffff05b2307 ()
Sep 26 15:48:06 ts genunix: [ID 802836 kern.notice] fffffe800085bbd0 fffffffff05824a6 ()
Sep 26 15:48:06 ts genunix: [ID 802836 kern.notice] fffffe800085bc10 fffffffff058279b ()
Sep 26 15:48:06 ts genunix: [ID 655072 kern.notice] fffffe800085bcf0 vboxdrv:supdrvIOCtl+d62 ()
Sep 26 15:48:06 ts genunix: [ID 655072 kern.notice] fffffe800085bdc0 vboxdrv:VBoxDrvSolarisIOCtl+359 ()
Sep 26 15:48:06 ts genunix: [ID 655072 kern.notice] fffffe800085bdd0 genunix:cdev_ioctl+1d ()
Sep 26 15:48:06 ts genunix: [ID 655072 kern.notice] fffffe800085bdf0 specfs:spec_ioctl+50 ()
Sep 26 15:48:06 ts genunix: [ID 655072 kern.notice] fffffe800085be20 genunix:fop_ioctl+25 ()
Sep 26 15:48:06 ts genunix: [ID 655072 kern.notice] fffffe800085bf00 genunix:ioctl+ac ()
Sep 26 15:48:06 ts genunix: [ID 655072 kern.notice] fffffe800085bf10 unix:brand_sys_syscall+21d ()
Sep 26 15:48:06 ts unix: [ID 100000 kern.notice]
Sep 26 15:48:06 ts genunix: [ID 672855 kern.notice] syncing file systems...
Sep 26 15:48:06 ts genunix: [ID 733762 kern.notice]  10
Sep 26 15:48:07 ts genunix: [ID 733762 kern.notice]  9
Sep 26 15:48:08 ts genunix: [ID 733762 kern.notice]  8
Sep 26 15:48:31 ts last message repeated 20 times
Sep 26 15:48:32 ts genunix: [ID 622722 kern.notice]  done (not all i/o completed)
Sep 26 15:48:33 ts genunix: [ID 111219 kern.notice] dumping to /dev/dsk/c2t0d0s1, offset 13746110464, content: kernel
Sep 26 15:48:46 ts genunix: [ID 409368 kern.notice] ^M100% done: 222774 pages dumped, compression ratio 4.44,
Sep 26 15:48:46 ts genunix: [ID 851671 kern.notice] dump succeeded

comment:21 follow-up: ↓ 22 Changed 5 years ago by ramshankar

Good, this driver should fix it:  http://www.virtualbox.org/download/testcase/vboxflt64.zip (amd64 only - not 32-bit)

To install: stop all VMs, and VBox GUI (check with ps -ef | grep -i vbox) to make sure no VBox process are running, then

as root:

rem_drv vboxflt
copy (vboxflt from the zip) to /platform/i86pc/kernel/drv/amd64/
add_drv vboxflt

And you're all set.

Once installed, modinfo | grep vb should show vboxflt now as 3.0.51r52856.

comment:22 in reply to: ↑ 21 Changed 5 years ago by half12

Replying to ramshankar:

Good, this driver should fix it:  http://www.virtualbox.org/download/testcase/vboxflt64.zip (amd64 only - not 32-bit)

To install: stop all VMs, and VBox GUI (check with ps -ef | grep -i vbox) to make sure no VBox process are running, then

as root:

rem_drv vboxflt
copy (vboxflt from the zip) to /platform/i86pc/kernel/drv/amd64/
add_drv vboxflt

And you're all set.

Hi,

Re-installed v3.0.7 last night and applied the vboxflt update.

Started VirtualBox this morning and attempted to recreate hang during moving VBox at Bios phase.

Started to VBox in uniprocessor mode, OpenSolaris b123 and WinXP while streaming radio on the Solaris 10 host. Within WinXP started playing a dvd full screen within the VBox. Within the OpenSolaris VBox I started the 30GB copy, this proceeded well and their was no audio disruption on the Solaris 10 host. However the dvd playback on WinXP kept on pausing briefly and then just stopped in fact the whole WinXP VBox had frozen, but the OpenSolaris VBox was working correctly.

comment:23 in reply to: ↑ 18 Changed 5 years ago by half12

Replying to half12:

Replying to ramshankar:

The time of reboot varies but it happens within a while, it's critical that this is a reboot and not the old hang while VMs were already running the system just hard froze all of a sudden.

Btw -  http://www.virtualbox.org/wiki/Core_dump <-- has instructions on how to get a core. In case of your reboot, if you had cores enabled I'm sure you'd have the core in /var/crash/<host>/ and the stack trace in "/var/adm/messages" file.

Thanks again for the testing!

Core Dumps were enabled

Do you want me to upload the vmcore file, details below.

# ls -l total 1816322 -rw-r--r-- 1 root root 2 Sep 26 15:54 bounds -rw-r--r-- 1 root root 2355728 Sep 26 15:51 unix.0 -rw-r--r-- 1 root root 927117312 Sep 26 15:54 vmcore.0

Messages error Sep 26 15:51:48 ts savecore: [ID 570001 auth.error] reboot after panic: BAD TRAP: type=e (#pf Page fault) rp=fffffe800085b250 addr=50 occurred in module "genunix" due to a NULL pointer dereference Sep 26 15:51:48 ts savecore: [ID 748169 auth.error] saving system crash dump in /var/crash/ts/*.0

comment:24 follow-up: ↓ 25 Changed 5 years ago by ramshankar

Are you 100% sure it's running the new driver?

Could you show me output of "modinfo | grep vb" and md5sum /platform/i86pc/kernel/drv/amd64/vboxflt and output of 'isainfo -k'

comment:25 in reply to: ↑ 24 Changed 5 years ago by half12

Replying to ramshankar:

Are you 100% sure it's running the new driver?

Could you show me output of "modinfo | grep vb" and md5sum /platform/i86pc/kernel/drv/amd64/vboxflt and output of 'isainfo -k'

# modinfo | grep vb
143 fffffffff0164810    a88 226   1  vboxnet (VirtualBox NetAdp 3.0.7r52830)
144 fffffffff03b1000   3650   -   1  vbi (VirtualBox Interfaces V6)
146 fffffffff03c0000  216d0 225   1  vboxdrv (VirtualBox HostDrv 3.0.7r52830)
204 fffffffff061f000   6a10 227   1  vboxflt (VirtualBox NetDrv 3.0.51r52884)
204 fffffffff061f000   6a10   -   1  vboxflt (VirtualBox NetMod 3.0.51r52884)

# digest -a md5 /platform/i86pc/kernel/drv/vboxflt
f13b7b612d9694523fb14c5761a425c0

# isainfo -k
amd64

comment:26 follow-up: ↓ 27 Changed 5 years ago by ramshankar

You posted md5sum of the 32-bit driver not the 64-bit that I requested (notice "amd64/vboxflt" in the path).

comment:27 in reply to: ↑ 26 Changed 5 years ago by half12

Replying to ramshankar:

You posted md5sum of the 32-bit driver not the 64-bit that I requested (notice "amd64/vboxflt" in the path).

Oops.

# digest -a md5 /platform/i86pc/kernel/drv/amd64/vboxflt d2d0f75c1a32a73c290dfc1bd33ad975

comment:28 follow-up: ↓ 29 Changed 5 years ago by ramshankar

That seems to be the on in the zip, could you post the stack trace from /var/adm/messages (the latest crash)?

comment:29 in reply to: ↑ 28 Changed 5 years ago by half12

Replying to ramshankar:

That seems to be the on in the zip, could you post the stack trace from /var/adm/messages (the latest crash)?

Nothing reported in /var/adm/messages and the problem only affected the WinXP Vbox only.

comment:30 Changed 5 years ago by ramshankar

Ah so this was not a full host freeze/hang. That means the issue you reported in  http://www.virtualbox.de/ticket/4950#comment:19 (comment 19) *IS* solved by the vboxflt driver and the core files you reported for this VM hang/crash is actually from the host reboot in comment #19.

For WinXP VM hang/crash, a VBox.log of the VM might be useful.

comment:31 follow-up: ↓ 33 Changed 5 years ago by ramshankar

I will close this defect as the host hang and reboots are solved. For the Win XP guest issue could you please open a new defect?

comment:32 Changed 5 years ago by ramshankar

  • Status changed from new to closed
  • Resolution set to fixed
  • Summary changed from VirtualBox v3.0.6 lock up with high network traffic to VirtualBox v3.0.6 lock up with high network traffic => fixed in SVN

comment:33 in reply to: ↑ 31 Changed 5 years ago by half12

Replying to ramshankar:

I will close this defect as the host hang and reboots are solved. For the Win XP guest issue could you please open a new defect?

Yes, as soon as I can reproduce it consistantly.

Note: See TracTickets for help on using tickets.

www.oracle.com
ContactPrivacy policyTerms of Use