Snap Creator on OVM

Hello, IHAC using OVM and would like to use SC for creating snapshots of linux VMs (system partitions). I have tried new SC-OVM plugin - works nicely, but during "xm pause" I loose 2-3 pings (or the response time is 5-9 s if no ping is lost), so it's unusable for production systems in 24x7 mode. Do you have any other idea, how to get extfs4 consistent snapshots of OS partitions (that is what the customer wants to "backup")? I can use RDM disks for every VM OS partition and create snapshot on controller using SSH from linux VM, before SSH I can call "sync; sync; sync" to empy as much cache as possible... Thanks a lot for advices in advance.Rostislav

we are actively investigating integration of Creator with NetBeans 5, but it's too early in the product development cycle to make any public statements about features or delivery dates
Rose

Similar Messages

CLI of snapcreator failed to access to the server

Hi,
I have installed the last version of snapcreator (version 4.1.1). But I can not use correctly the CLI command snapcreator. Its make an error.
Two examples :root@hefrscs01:~# /opt/snapcreator/scServer4.1.1/snapcreator --server snapcreator.XXX.XX --port 8443 --user XXXXX --passwd XXXXX--action pingAgent --agentName agetestdb1.XXXXX.XX --agentPort 9090 --verbose500 Can't connect to snapcreator.hefr.ch:8443 at /</opt/snapcreator/scServer4.1.1/snapcreator>SnapCreator/Service/Engine.pm line 58
root@hefrscs01:~# /opt/snapcreator/scServer4.1.1/snapcreator --server snapcreator.XXX.XX --port 8443 --user XXXXX --passwd XXXXX --profile pr_agetestdb1 --config DBA01_backup_data --action backupList --policy DAILY --verbose500 Can't connect to snapcreator.hefr.ch:8443 at /</opt/snapcreator/scServer4.1.1/snapcreator>SnapCreator/Service/Repository.pm line 50root@hefrscs01:~#
I need some help to resovle this error.
Regards/Olivier

Hi, check this link, it seems to be the same problem with openssl:http://community.netapp.com/t5/Snap-Creator-Framework-Discussions/SC-4-1C-CLI-issue/td-p/7465

Netapp Plug-In for Symantec NetBackup and SMTape

I am trying to provide a solution to a customer that involves SMTape to seed SnapMirror (VSM) relationships across a slow network connection between TX and KY. I've found some really good docs about doing this. However, I would like to be able to leverage the customer's existing Symantec NetBackup and tape library solution to avoid having to procure/connect new tape drives to the six (6) storage controllers in this environment. Is it possible to use NetBackup to trigger SMTape? So far I'm not seeing "YES" in any of the documentation so before I assume "NO" I wanted to bring it up in the community... Perhaps someone out there has solved this for their customer or can provide some insight as to how best to deploy SMTape for this situation.Ontap 8.2P3 (7-mode)FAS3250 HA pair in TX (PROD)FAS6220 HA pair in TX (DEV)FAS3250 HA pair in KY (DR)Symantec NetBackup v7.3Thank you!Greg

Hi GregHi Ahmad I'm also interesting in doing a SMTape using NetBackup. It should be used to get a LPAR Filesystem ( no database) to tape. All should be managed using Snap Creator. Here my theoretical stepsSnap Creator triggers LAPR to get file systems consistent via pre commandSnap Creator triggers snap shotSnap Creator triggers NetBackup to run job via post command netbackup/bin/bpbackup -i -p policy_name -s schedule_name -S....The specifiy NetBackup policy is careate with the set type - smtape option and the whole source vol is selected. Currently this job failes with "Error ndmpagent(pid=3144) source servername: DATA&colon; Backup terminated: backup configuration failure (2) on RPC error: ERROR: BAD ENV OP (for /vol/volname)" Any idea?Furthermore I want to have used the snapshot created by Snap Creator for the smtape process. I've see that there are environmen variables available e. g. SMTAPE_SNAPSHOT_NAME but how I have to use them? And is there any NetApp licensing needed? In document it says no but the NetApp option snapmirror.enable can only be set to on with a vailed license. Thanks for any feeback Thorsten

APP_IGNORE_ERROR to Y, but no Warning anymore in SC 4.1.1

Hi All In SC 3.5 we have set the Option APP_IGNORE_ERROR to Y, because we have a lot of DBs running on one Server. When one DB Backup fails the Job doesn't stop (because APP_IGNORE_ERROR=Y), but the Job went yellow in the Job Monitor. This Yellow Indicator is important for us for checking our Backups. Now in 4.1.1P4D3 we have also set the Option APP_IGNORE_ERROR to Y, but when one DB Backup fails the Job still went green in the Job Monitor. Is this a BUG or a wanted behaviour? TIAThomas

Hi John Thanks for your answer. We use SC 4.1.1P4D3. I have already open a Case.In the Config File I found following explanation: # APP_IGNORE_ERROR - (optional) - (Y|N) ## If set to Y will cause Snap Creator to not exit when encountering an application error. An error message will # be sent if SENDTRAP is configured but Snap Creator will not exit. This may be useful if you are backing up# multiple databases and dont want a single database failure to stop the backup So In my Understanding, SC should generate a Error (Warning) Message at least via sendtrap. I will Update this thread, when I have further Info from Netapp Support regardsThomas

OVM 3.0.3 import VirtualBox OVM Manager 3.0.3 OVA template

VirtualBox Template for Oracle VM Manager 3.0.3,
http://www.oracle.com/technetwork/server-storage/vm/template-1482544.html
The OVM 3.0.3 host should be configured with enough space to accommodate import of OVA template. In
this case 36G was allocated for "sda",
[root@ovm303 ~]# fdisk -l /dev/sda
Disk /dev/sda: 38.6 GB, 38654705664 bytes
255 heads, 63 sectors/track, 4699 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System
/dev/sda1 * 1 4189 33648111 83 Linux
/dev/sda2 4190 4699 4096575 82 Linux swap / Solaris
Post OVA import disk usage,
[root@ovm303 ~]# df -T -t ext3
Filesystem Type 1K-blocks Used Available Use% Mounted on
/dev/sda1 ext3 32594216 24105604 6806208 78% /.
[root@ovm303 ~]# cat /etc/motd
Warning: making manual modifications in the management domain
might cause inconsistencies between Oracle VM Manager and the server.
[root@ovm303 ~]# head -4 /etc/ovs-info
OVS summary
release: 3.0.3
date: 201111300901
build: 123
The contents of "ovs-info" also include package information.
The "vbox-img" package is something new in OVM 3.0.3,
[root@ovm303 ~]# grep vbox-img /etc/ovs-info
vbox-img-1.0-2
[root@ovm303 ~]# rpm -ql vbox-img
/usr/bin/vbox-img
/usr/share/doc/vbox-img-1.0
/usr/share/doc/vbox-img-1.0/COPYING
[root@ovm303 ~]# vbox-img --help
Oracle VM VirtualBox Disk Utility 4.1.3
(C) 2005-2011 Oracle Corporation
All rights reserved.
Usage: vbox-img
setuuid --filename <filename>
[--format VDI|VMDK|VHD|...]
[--uuid <uuid>]
[--parentuuid <uuid>]
[--zeroparentuuid]
convert --srcfilename <filename>
--dstfilename <filename>
[--stdin]|[--stdout]
[--srcformat VDI|VMDK|VHD|RAW|..]
[--dstformat VDI|VMDK|VHD|RAW|..]
[--variant Standard,Fixed,Split2G,Stream,ESX]
info --filename <filename>
compact --filename <filename>
createcache --filename <filename>
--size <cache size>
Download "OracleVMManager3" OVA,
[root@ovm303 xen]# mkdir /etc/xen/ovm303-m
[root@ovm303 ~]# cd /etc/xen/ovm303-m
[root@ovm303 xen]# wget http://download.oracle.com/otn-pub/otn_software/OracleVMManager3.ova
--2012-03-14 17:06:01-- http://download.oracle.com/otn-pub/otn_software/OracleVMManager3.ova
Length: 4234452992 (3.9G) [text/plain]
Saving to: `OracleVMManager3.ova'
100%[====================================>] 4,234,452,992 834K/s in 55m 36s
2012-03-14 18:01:38 (1.21 MB/s) - `OracleVMManager3.ova' saved [4234452992/4234452992]
[root@ovm303 xen]# md5sum OracleVMManager3.ova
75fc334422eda6c3425a2f09fb88f9e0 OracleVMManager3.ova
Unpack OVA,
[root@ovm303 xen]# tar tvf OracleVMManager3.ova
-rw------- someone/someone 59626 2012-01-12 17:42:49 OracleVMManager3.ovf
-rw------- someone/someone 1987182592 2012-01-12 17:48:06 OracleVMManager3-disk1.vmdk
-rw------- someone/someone 2247208960 2012-01-12 17:51:24 OracleVMManager3-disk2.vmdk
[root@ovm303 ovm303-m]# tar xvf OracleVMManager3.ova
OracleVMManager3.ovf
OracleVMManager3-disk1.vmdk
OracleVMManager3-disk2.vmdk
Use "vbox-img" to show "OracleVMManager3-disk1.vmdk",
[root@ovm303 ovm303-m]# vbox-img info --filename OracleVMManager3-disk1.vmdk
--- Dumping VD Disk, Images=1
Dumping VD image "OracleVMManager3-disk1.vmdk" (Backend=VMDK)
Header: Geometry PCHS=16383/16/63 LCHS=1024/255/63 cbSector=52428800
Header: uuidCreation={f3859e00-962c-440f-a63d-feb7b731d5da}
Header: uuidModification={00000000-0000-0000-0000-000000000000}
Header: uuidParent={00000000-0000-0000-0000-000000000000}
Header: uuidParentModification={00000000-0000-0000-0000-000000000000}
Convert "OracleVMManager3-disk1.vmdk" to RAW,
[root@ovm303 ovm303-m]# vbox-img convert srcfilename OracleVMManager3-disk1.vmdk dstfilename OracleVMManager3-disk1.img srcformat VMDK dstformat RAW
Converting image "OracleVMManager3-disk1.vmdk" with size 26843545600 bytes (25600MB)...
[root@ovm303 ovm303-m]# file OracleVMManager3-disk1.img
OracleVMManager3-disk1.img: x86 boot sector; partition 1: ID=0x83, active, starthead 1, startsector 63, 208782 sectors; partition 2: ID=0x8e, starthead 0, startsector 208845, 52211250 sectors, code offset 0x48
[root@ovm303 ovm303-m]# fdisk -l OracleVMManager3-disk1.img
last_lba(): I don't know how to handle files with mode 8180
You must set cylinders.
You can do this from the extra functions menu.
Disk OracleVMManager3-disk1.img: 0 MB, 0 bytes
255 heads, 63 sectors/track, 0 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System
OracleVMManager3-disk1.img1 * 1 13 104391 83 Linux
OracleVMManager3-disk1.img2 14 3263 26105625 8e Linux LVM
Partition 2 has different physical/logical endings:
phys=(1023, 254, 63) logical=(3262, 254, 63)
Use "vbox-img" to show "OracleVMManager3-disk2.vmdk",
[root@ovm303 ovm303-m]# vbox-img info --filename OracleVMManager3-disk2.vmdk
--- Dumping VD Disk, Images=1
Dumping VD image "OracleVMManager3-disk2.vmdk" (Backend=VMDK)
Header: Geometry PCHS=16383/16/63 LCHS=1024/255/63 cbSector=20971520
Header: uuidCreation={abc43be0-7a43-4ba2-a205-eaf37184a4d3}
Header: uuidModification={00000000-0000-0000-0000-000000000000}
Header: uuidParent={00000000-0000-0000-0000-000000000000}
Header: uuidParentModification={00000000-0000-0000-0000-000000000000}
Convert "OracleVMManager3-disk2.vmdk" to RAW,
[root@ovm303 ovm303-m]# vbox-img convert srcfilename OracleVMManager3-disk2.vmdk dstfilename OracleVMManager3-disk2.img srcformat VMDK dstformat RAW
Converting image "OracleVMManager3-disk2.vmdk" with size 10737418240 bytes (10240MB)...
[root@ovm303 ovm303-m]# file OracleVMManager3-disk2.img
OracleVMManager3-disk2.img: x86 boot sector; partition 1: ID=0x83, starthead 1, startsector 63, 20964762 sectors, extended partition table (last)\011
[root@ovm303 ovm303-m]# fdisk -l OracleVMManager3-disk2.img
last_lba(): I don't know how to handle files with mode 8180
You must set cylinders.
You can do this from the extra functions menu.
Disk OracleVMManager3-disk2.img: 0 MB, 0 bytes
255 heads, 63 sectors/track, 0 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System
OracleVMManager3-disk2.img1 1 1305 10482381 83 Linux
Partition 1 has different physical/logical endings:
phys=(1023, 254, 63) logical=(1304, 254, 63)
Update disk image "/boot" configuration using loopback mount,
[root@ovm303 ovm303-m]# mkdir boot
[root@ovm303 ovm303-m]# kpartx -l OracleVMManager3-disk1.img
loop1p1 : 0 208782 /dev/loop1 63
loop1p2 : 0 52211250 /dev/loop1 208845
loop deleted : /dev/loop1
[root@ovm303 ovm303-m]# kpartx -a OracleVMManager3-disk1.img
[root@ovm303 ovm303-m]# mount /dev/mapper/loop1p1 boot
Update "/boot/grub/device.map",
[root@ovm303 ovm303-m]# cat boot/grub/device.map
# this device map was generated by anaconda
(hd0) /dev/hda
Update "/boot/grub/grub.conf" to use non-UEK kernel to avoid issues with OL5u7 default UEK boot and HVM,
[root@ovm303 ovm303-m]# grep default boot/grub/grub.conf
default=1
Mount "/" logical volume,
[root@ovm303 ovm303-m]# vgscan
Reading all physical volumes. This may take a while...
Found volume group "VolGroup00" using metadata type lvm2
[root@ovm303 ovm303-m]# vgchange -ay VolGroup00
2 logical volume(s) in volume group "VolGroup00" now active
[root@ovm303 ovm303-m]# lvs
LV VG Attr LSize Origin Snap% Move Log Copy% Convert
LogVol00 VolGroup00 -wi-a- 19.06G
LogVol01 VolGroup00 -wi-a- 5.81G
[root@ovm303 ovm303-m]# mkdir root
[root@ovm303 ovm303-m]# ls /dev/VolGroup00/
LogVol00 LogVol01
[root@ovm303 ovm303-m]# mount /dev/VolGroup00/LogVol00 root
Update "/etc/inittab" to boot single user,
[root@ovm303 ovm303-m]# grep id root/etc/inittab
id:1:initdefault:
Update "/etc/fstab" to ignore "/u01" block device mount,
[root@ovm303 ovm303-m]# grep sdb root/etc/fstab
#/dev/sdb1 /u01 ext3 defaults 1 1
Unmount "/boot" and "/",
[root@ovm303 ovm303-m]# umount boot
[root@ovm303 ovm303-m]# umount root
[root@ovm303 ovm303-m]# vgchange -an VolGroup00
0 logical volume(s) in volume group "VolGroup00" now active
[root@ovm303 ovm303-m]# kpartx -d OracleVMManager3-disk1.img
loop deleted : /dev/loop1
The OVM 3.0.3 network bridge configuration uses "bond0" vs "xenbr0".
Configure "vm.cfg",
[root@ovm303 ovm303-m]# cat vm.cfg
# (H)ardware (V)irtual (M)achine configuration
# /etc/xen/xmexample.hvm
kernel = 'hvmloader'
builder = 'hvm'
memory = '4096'
# The number of cpus guest platform has, default=1
#vcpus=1
# PXE boot
#bootloader = '/usr/bin/pypxeboot'
#bootloader = '/usr/bin/pygrub'
# Virtual block device configuration using file:
# dd if=/dev/zero of=/$VM_PATH/$VM_NAME/System.img bs=1M count=$MB
# disk = [ 'file:/$VM_PATH/$VM_NAME/System.img,ioemu:hda,w', ]
# Configuration using physical disk logical volume:
# lvcreate --size ${GB}G --name $VM_NAME $VG_NAME
# disk = [ 'phy:/dev/$VG_NAME/$VM_NAME,xvda,w',]
# CDROM configuration using file:
# disk = [ 'file:/$BOOT_ISO_NAME,hdc:cdrom,r', ]
disk = [ 'file:/etc/xen/ovm303-m/OracleVMManager3-disk1.img,ioemu:hda,w', 'file:/etc/xen/ovm303-m/OracleVMManager3-disk2.img,ioemu:hdb,w', ]
# Domain name
name = "ovm303-m"
# Networking
# Network bridge status:
# /etc/xen/scripts/network-bridge status
# 'bridge' should be proper interface e.g. xenbr0
vif = ['type=ioemu, bridge=bond0']
# boot on floppy (a), hard disk (c), Network (n) or CD-ROM (d)
# default: hard disk, cd-rom, floppy
boot = "cdn"
# Enable HV VNC console
# Uses avaiablle next VNC port $OVM_SERVER_HOST:0 == 5900, :1 == 5901 etc
# Use "xm list -l" to determince VNC port "location"
vnc=1
vncunused=1
vnclisten="0.0.0.0"
vncpasswd=''
sdl=0
# Set the real time clock to local time [default=0 i.e. set to utc]
# Allows Win VM clock to work with UTC on OVM host system.
#localtime=1
Start VM,
[root@ovm303 ovm303-m]# xm create vm.cfg
Using config file "./vm.cfg".
Error: Device 0 (vif) could not be connected. Hotplug scripts not working.
Current bridge,
[root@ovm303 ~]# brctl show
bridge name bridge id STP enabled interfaces
Update "/etc/xen/xend-config.sxp" to enable "network-bridge",
[root@ovm303 ovm303-m]# cp /etc/xen/xend-config.sxp /etc/xen/xend-config.sxp.orig
[root@ovm303 ovm303-m]# diff /etc/xen/xend-config.sxp /etc/xen/xend-config.sxp.orig
165,167d164
< # Enable network-bridge, JAP 20120322
< (network-script network-bridge)
<
[root@ovm303 ~]# service xend restart
Restarting xend...
xend daemon (pid 8848) is running...
[root@ovm303 ~]# brctl show
bridge name bridge id STP enabled interfaces
bond0 8000.002128d5ce86 no pbond0
Start VM,
[root@ovm303 ovm303-m]# xm create vm.cfg
Using config file "./vm.cfg".
Started domain ovm303-m (id=1)
Determine VNC console session,
[root@ovm303 ovm303-m]# xm list -l ovm303-m | grep location
(location 0.0.0.0:5900)
(location 3)
VNC console session. The "oraclevm-template" tools are not run in single user so hostname is "localhost",
[root@localhost ~]# who -r
run-level S 2012-03-23 09:09 last=1
Update "/etc/fstab" for "/u01" mount to use label,
[root@localhost ~]# grep /u01 /etc/fstab
#/dev/sdb1 /u01 ext3 defaults 1 1
[root@localhost ~]# blkid
/dev/mapper/VolGroup00-LogVol01: TYPE="swap"
/dev/mapper/VolGroup00-LogVol00: UUID="2751f1b2-836b-4786-ae71-8b402e0f8503" TYPE="ext3"
/dev/VolGroup00/LogVol00: UUID="2751f1b2-836b-4786-ae71-8b402e0f8503" SEC_TYPE="ext2" TYPE="ext3"
/dev/VolGroup00/LogVol01: TYPE="swap"
/dev/hda1: LABEL="/boot" UUID="8701866e-5ed5-49f7-9597-bb9709d22359" TYPE="ext3"
/dev/hdb1: UUID="5f9a7bef-3a6d-4a83-b288-638537424fa7" SEC_TYPE="ext2" TYPE="ext3"
[root@localhost ~]# e2label /dev/hdb1 u01
[root@localhost ~]# grep u01 /etc/fstab
LABEL=u01 /u01 ext3 defaults 1 1
[root@localhost ~]# mount /u01
[root@localhost ~]# df /u01
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/hdb1 10317828 4250832 5542880 44% /u01
As configured the default run level will be elevated to "5" by the "/etc/rc.local" script. This enables a
console desktop for user "ovm". This is unneeded and will be implemented by VNC session for user
"ovm".
Update "/etc/inittab" to enable full multiuser mode run level 3,
[root@localhost ~]# grep id /etc/inittab
id:3:initdefault:
Disable "/etc/rc.local" run level elevation,
[root@localhost ~]# tail -2 /etc/rc.local
# Disable local forced runlevel 5, JAP 20120404
#telinit 5
Disable firewall to allow VNC connections,
[root@localhost ~]# chkconfig --list iptables
iptables 0:off 1:off 2:on 3:on 4:on 5:on 6:off
[root@localhost ~]# chkconfig iptables off
[root@localhost ~]# chkconfig --list ip6tables
ip6tables 0:off 1:off 2:on 3:on 4:on 5:on 6:off
[root@localhost ~]# chkconfig ip6tables off
Reboot,
[root@localhost ~]# reboot
Broadcast message from root (pts/0) (Mon Apr 2 09:22:17 2012):
The system is going down for reboot NOW!
Post "oraclevm-template" configuration e.g. defining FQDN and static IP addresss etc. OVM Manager 3.0.3 is up.
Configure VNC for "ovm",
[root@ovm303-m ~]# su - ovm
[ovm@ovm303-m ~]$ vncserver
You will require a password to access your desktops.
Password: Welcome1
Verify: Welcome1
xauth: creating new authority file /home/ovm/.Xauthority
New 'ovm303-m:1 (ovm)' desktop is ovm303-m:1
Creating default startup script /home/ovm/.vnc/xstartup
Starting applications specified in /home/ovm/.vnc/xstartup
Log file is /home/ovm/.vnc/ovm303-m:1.log
[ovm@ovm303-m ~]$ vncserver -kill :1
Killing Xvnc process ID 12499
Update "xstartup",
[ovm@ovm303-m ~]$ cat /home/ovm/.vnc/xstartup
#!/bin/sh
# Uncomment the following two lines for normal desktop:
unset SESSION_MANAGER
exec /etc/X11/xinit/xinitrc
#[ -x /etc/vnc/xstartup ] && exec /etc/vnc/xstartup
#[ -r $HOME/.Xresources ] && xrdb $HOME/.Xresources
#xsetroot -solid grey
#vncconfig -iconic &
#xterm -geometry 80x24+10+10 -ls -title "$VNCDESKTOP Desktop" &
#twm &
[ovm@ovm303-m ~]$ vncserver
New 'ovm303-m:1 (ovm)' desktop is ovm303-m :1
Starting applications specified in /home/ovm/.vnc/xstartup
Log file is /home/ovm/.vnc/ovm303-m:1.log
Configure "/etc/sysconfig/vncservers",
[root@ovm303-m ~]# chkconfig --list vncserver
vncserver 0:off 1:off 2:off 3:off 4:off 5:off 6:off
[root@ovm303-m ~]# chkconfig vncserver on
[root@ovm303-m ~]# chkconfig --list vncserver
vncserver 0:off 1:off 2:on 3:on 4:on 5:on 6:off
[root@ovm303-m ~]# tail -2 /etc/sysconfig/vncservers
# Start VNC session for ovm, JAP 20120404
VNCSERVERS="1:ovm"

Thanks Jim - looks great!

How can i work without grid in sun creator

Hello guys,
how can i place component in sun creator anywhere i like instead of being restricted to place it within the grid (i.e placing component within the box drawn in sun creator (specific x and y axis))
thanks
Message was edited by:
kaskalee

Go to Tools|Options and select Visual Designer.
Clear Snap to Grid checkbox. This should let you precisely place a component anywhere on the canvas.
If you don't want the grid lines, clear the Show Grid checkbox.

OVM support for NPIV and pass through disks

Hi Gurus,
I have a couple of questions -
1) Does OVM support NPIV?
2) Does mapping the physical disks via the phy: definition in the vm.cfg file equivalent of pass through disks to the guest OS?.
thanks!
Gak

avasek wrote:
Thanks!. The pass through question was actually in context of usage with snap* products from netapp as they work with VMware. All the snap* work with RDM's in VM's (pass through's), so i was wondering if the snap* could be made to work with the phy: def to the VM's as well. The RDM def should be equivalent of the phy: def ..right?. So ideally, they ought to work?No idea, to be honest -- I haven't worked with NetApp gear yet. You're probably better off checking with NetApp directly as well.

How can I convert an OVM vm (.img) to VDI (virtualbox)?

Hi,
how can I convert an OVM vm (.img) to VDI (virtualbox)?
Is .img format a raw image? If yes can I use "VBoxManage convertdd" to convert from .img to vdi?
tnx & regards,
S.

Hi,
"VBoxManage convertdd" solved.
Regards,
S.

Java API, problem with secondary keys using multi key creator

Hi,
I'm developing an application that inserts a few 100k or so records into a Queue DB, then access them using one of four Hash SecondaryDatabases. Three of these secondary dbs use a SecondaryMultiKeyCreator to generate the keys, and one uses a SecondaryKeyCreator. As a test, I'm trying to iterate through each secondary key. When trying to iterate through the keys of any of the secondary databases that use a SecondaryMultiKeyCreator, I have problems. I'm attempting to iterate through the keys by:
1: creating a StoredMap of the SecondaryDatabase
2: getting a StoredKeySet from said map
3: getting a StoredIterator on said StoredKeySet
4: iterate
The first call to StoredIterator.next() fails at my key binding (TupleBinding) because it is only receiving 2 bytes of data for the key, when it should be getting more, so an Exception is thrown. I suspected a problem with my SecondaryMultiKeyCreator, so I added some debug code to check the size of the DatabaseEntries it creates immediately before adding them to the results key Set. It checks out right. I also checked my key binding like so:
1: use binding to convert the key object to a DatabaseEntry
2: use binding to convert the created DatabaseEntry back to a key object
3: check to see if the old object contains the same data as the new one
Everything checked out ok.
What it boils down to is this: my key creator is adding DatabaseEntries of the correct size to the results set, but when the keys are being read back later on, my key binding is only receiving 2 bytes of data. For the one SecondaryDatabase that doesn't use a SecondaryMultiKeyCreator, but just a SecondaryKeyCreator, there are no issues and I am able to iterate through its secondary keys as expected.
EDIT: New discovery: if I only add ONE DatabaseEntry to the results set in my SecondaryMultiKeyCreator, I am able to iterate through the keys as I would like to.
Any ideas or suggestions?
Thank you for your attention,
-Justin
Message was edited by:
Steamroller

Hi Justin,
Sorry about the delayed response here. I have created a patch that resolves the problem.
If you apply the patch to your 4.6.21 source tree, and then rebuild Berkeley DB, the improper behavior should be resolved.
Regards,
Alex
diff -rc db/db_am.c db/db_am.c
*** db/db_am.c     Thu Jun 14 04:21:30 2007
--- db/db_am.c     Fri Jun 13 11:20:28 2008
*** 331,338 ****
     F_SET(dbc, DBC_TRANSIENT);
!      switch (flags) {
!      case DB_APPEND:
           * If there is an append callback, the value stored in
           * data->data may be replaced and then freed. To avoid
--- 331,337 ----
     F_SET(dbc, DBC_TRANSIENT);
!      if (flags == DB_APPEND && LIST_FIRST(&dbp->s_secondaries) == NULL) {
           * If there is an append callback, the value stored in
           * data->data may be replaced and then freed. To avoid
*** 367,388 ****
-            * Secondary indices: since we've returned zero from an append
-            * function, we've just put a record, and done so outside
-            * __dbc_put. We know we're not a secondary-- the interface
-            * prevents puts on them--but we may be a primary. If so,
-            * update our secondary indices appropriately.
-            * If the application is managing this key's data, we need a
-            * copy of it here. It will be freed in __db_put_pp.
-           DB_ASSERT(dbenv, !F_ISSET(dbp, DB_AM_SECONDARY));
-           if (LIST_FIRST(&dbp->s_secondaries) != NULL &&
-               (ret = __dbt_usercopy(dbenv, key)) == 0)
-                ret = __db_append_primary(dbc, key, &tdata);
           * The append callback, if one exists, may have allocated
           * a new tdata.data buffer. If so, free it.
--- 366,371 ----
*** 390,401 ****
          /* No need for a cursor put; we're done. */
          goto done;
!      default:
!           /* Fall through to normal cursor put. */
!           break;
!      if (ret == 0)
          ret = __dbc_put(dbc,
              key, data, flags == 0 ? DB_KEYLAST : flags);
--- 373,379 ----
          /* No need for a cursor put; we're done. */
          goto done;
!      } else
          ret = __dbc_put(dbc,
              key, data, flags == 0 ? DB_KEYLAST : flags);
diff -rc db/db_cam.c db/db_cam.c
*** db/db_cam.c     Tue Jun 5 21:46:24 2007
--- db/db_cam.c     Thu Jun 12 16:41:29 2008
*** 899,905 ****
     DB_ENV *dbenv;
     DB dbp, sdbp;
     DBC dbc_n, oldopd, opd, sdbc, *pdbc;
!      DBT olddata, oldpkey, newdata, pkey, temppkey, tempskey;
     DBT all_skeys, skeyp, *tskeyp;
     db_pgno_t pgno;
     int cmp, have_oldrec, ispartial, nodel, re_pad, ret, s_count, t_ret;
--- 899,905 ----
     DB_ENV *dbenv;
     DB dbp, sdbp;
     DBC dbc_n, oldopd, opd, sdbc, *pdbc;
!      DBT olddata, oldpkey, newdata, pkey, temppkey, tempskey, tdata;
     DBT all_skeys, skeyp, *tskeyp;
     db_pgno_t pgno;
     int cmp, have_oldrec, ispartial, nodel, re_pad, ret, s_count, t_ret;
*** 1019,1026 ****
      * should have been caught by the checking routine, but
      * add a sprinkling of paranoia.
!      DB_ASSERT(dbenv, flags == DB_CURRENT || flags == DB_KEYFIRST ||
!            flags == DB_KEYLAST || flags == DB_NOOVERWRITE);
      * We'll want to use DB_RMW in a few places, but it's only legal
--- 1019,1027 ----
      * should have been caught by the checking routine, but
      * add a sprinkling of paranoia.
!      DB_ASSERT(dbenv, flags == DB_APPEND || flags == DB_CURRENT ||
!            flags == DB_KEYFIRST || flags == DB_KEYLAST ||
!            flags == DB_NOOVERWRITE);
      * We'll want to use DB_RMW in a few places, but it's only legal
*** 1048,1053 ****
--- 1049,1107 ----
               goto err;
          have_oldrec = 1; /* We've looked for the old record. */
+      } else if (flags == DB_APPEND) {
+           /*
+            * With DB_APPEND, we need to do the insert to populate the
+            * key value. So we swap the 'normal' order of updating
+            * secondary / verifying foreign databases and inserting.
+            *
+            * If there is an append callback, the value stored in
+            * data->data may be replaced and then freed. To avoid
+            * passing a freed pointer back to the user, just operate
+            * on a copy of the data DBT.
+            */
+           tdata = *data;
+           /*
+            * If this cursor is going to be closed immediately, we don't
+            * need to take precautions to clean it up on error.
+            */
+           if (F_ISSET(dbc_arg, DBC_TRANSIENT))
+                dbc_n = dbc_arg;
+           else if ((ret = __dbc_idup(dbc_arg, &dbc_n, 0)) != 0)
+                goto err;
+
+           pgno = PGNO_INVALID;
+
+           /*
+            * Append isn't a normal put operation; call the appropriate
+            * access method's append function.
+            */
+           switch (dbp->type) {
+           case DB_QUEUE:
+                if ((ret = __qam_append(dbc_n, key, &tdata)) != 0)
+                     goto err;
+                break;
+           case DB_RECNO:
+                if ((ret = __ram_append(dbc_n, key, &tdata)) != 0)
+                     goto err;
+                break;
+           default:
+                /* The interface should prevent this. */
+                DB_ASSERT(dbenv,
+                    dbp->type == DB_QUEUE || dbp->type == DB_RECNO);
+
+                ret = __db_ferr(dbenv, "DBC->put", 0);
+                goto err;
+           }
+           /*
+            * The append callback, if one exists, may have allocated
+            * a new tdata.data buffer. If so, free it.
+            */
+           FREE_IF_NEEDED(dbenv, &tdata);
+           pkey.data = key->data;
+           pkey.size = key->size;
+           /* An append cannot be replacing an existing item. */
+           nodel = 1;
     } else {
          /* Set pkey so we can use &pkey everywhere instead of key. */
          pkey.data = key->data;
*** 1400,1405 ****
--- 1454,1465 ----
skip_s_update:
+       * If this is an append operation, the insert was done prior to the
+       * secondary updates, so we are finished.
+       */
+      if (flags == DB_APPEND)
+           goto done;
+      /*
      * If we have an off-page duplicates cursor, and the operation applies
      * to it, perform the operation. Duplicate the cursor and call the
      * underlying function.

How to configure multiple IP-Adresses on one NIC in OVM

Hi guys,
we have some OVM Servers 2.2.1 installed. They all have three NICs in them (eth0-2). For eth0 and eth2 we need multiple IP-Adresses configured to work with our storage solution. The configuration should look like this:
eth0:0 - 192.168.31.x
eth0:1 - 192.168.33.x
eth1 - 192.168.0.x
eth2:0 - 192.168.32.x
eth2:1 - 192.168.34.x
I tried creating gonfig files ifcfg-eth0:0 in /etc/sysconfig/network-scripts with the following content
DEVICE=eth0:0
BOOTPROTO=static
ONBOOT=yes
NETMASK=255.255.255.0
IPADDR=192.168.31.3
TYPE=Ethernet
HWADDR=00:1b:21:06:f0:fa
But it does not work. Neither eth0:0 nore an additional xenbr is configured when i restart.
I can configure multiple ip addresses for xenbr0 by calling something like "ifconfig xenbr0 192.168.31.x". But these settings get lost as sonn as i reboot the server.
I have already searched google, forums and blogs for a solution but could not find one. I thought that this should be a pretty common task. In "normal" linux it is not an issue at all.
Has somebody any experience with a setup like this? I would really apreciate your help.
Thanks
Marek

Hi,
configuring IP aliasses seems to work the same way like configuring VLANs. You will have to configure the appropriate bridges at system startup manually.
E.g. this little script is what I use to configure VLANs and briges on my OVM:
#!/bin/bash
echo "Configuring VLANs and bridges\n"
echo "Configuring VLAN 24 on eth2"
vconfig add eth2 24
brctl addbr br2.24
brctl addif br2.24 eth2.24
ifconfig eth2.24 up
ifconfig br2.24 up
echo "Configuring VLAN 31 on eth2"
vconfig add eth2 31
brctl addbr br2.31
brctl addif br2.31 eth2.31
ifconfig eth2.31 up
ifconfig br2.31 up
This script is run at startup by init using this startup skript in /etc/init.d, which is what I think you're missing as you can already seem to bring up the needed IPs manually - you'd have to change the chkconfig accordingly, that is make sure that the config script gets startet after the network has been brought up, like this: # chkconfig: 2345 11 91
[root@oraclevm01 ~]# cat /etc/init.d/vlan-config
#! /bin/bash
# vlan-config configure VLANs
# chkconfig: 2345 11 89
# description: Configures VLANs at boot time before the \
# network is started.
### BEGIN INIT INFO
# Provides: $vlan
### END INIT INFO
# Source function library.
. /etc/init.d/functions
if [ ! -f /etc/vlan-config ]; then
exit 0
fi
# See how we were called.
case $1 in
     start)
               # configure vlans
               /etc/vlan-config
          echo $"Usage: $0 {start}"
          exit 1
esac
exit 0
Hth,
budy
Edited by: budachst on Jun 29, 2011 8:11 AM

Satellite U400 Impossible to create recovery DVD with Recovery disc creator

Hi,
I have a Satellite U400 (few years old), and I never created the recovery DVD.
Now I want to do it, but for some reasons, when I launch Toshiba recovery disc creator, nothing happens...
The partition with the recovery files is there, as well as the folder HDDRecovery. I did add some files to this partition, could this be the problem ?
Thanks for your help.

It is big problem now if the Toshiba tool cannot find the files.
How it works you can read on http://forums.computers.toshiba-europe.com/forums/thread.jspa?messageID=140985
Check it out please.
What you can try now is to check if HDD recovery installation will start - http://aps2.toshiba-tro.de/kb0/HTD1303440001R01.htm but Im afraid the same problem will occurs again.
To be honest you should install Win7 32bit on this small Satellite notebook. It works perfectly. Forget original Vista OS, Win7 works better.

MS Access and Sun Java Studio Creator

I am new to the forum.
Having searched for the issue above I failed to find a way to connect between MS Access database and Creator.
However I came accross a thread in 2004 which suggested that Sun techs are working on a solution.
I wonder if such a solution has been found ?
If not I may have to convert my database to SQL to be able to use it with Creator or consider my purchase of Creator was a loss.
This may be a better long term but very sad for someone who is using Creator for developing purposes.
I will value your help.

I also would find a driver for MS Access very useful. I currently have about 12 databases done in MS Access that we are trying to move over to something else. We are evaluating Creator for this and it would be useful if I could talk to the existing databases with it.
We could port the databases over to SQL server but then I have to load that onto my home computer because I do a lot of work there.
I read somewhere that it might be possible to use the JDBC-ODBC data bridge to read the access files but I am not sure how to set that up. I can find several 3rd party drivers but I woudl for a non-profit and there is never any money to buy stuff like that .

How do i take a snap shot on my mac?, how do i take a snap shot on my mac?

how do i take a snap shot on my mac?, how do i take a snap shot on my mac?

Mac OS X 10.7 Help: Shortcuts for taking pictures of the screen

[installation] studio creator on Suse 10.1

Hi there,
on my laptop (1,7 Ghz cpu, 1 Ghz ram), runs Suse 10.1(kernel Linux 2.6.16.13-4-default i686) and I'd like install studio creator wich is at moment on XP (and I'm not very happy about the performance).
But looking at http://developers.sun.com/prodtech/javatools/jscreator/sysreqs/index.jsp#04
it seems that it can only runs on Red Hat Enterprise Linux 3.
Did anyone try it on Suse, what will be the conseguence if I proceed anyway?
Thanks in advance for your reply.

Have you tried downloading and installing the latest HotFixes from the Update Center, on your window machine? The HotFixes have several performance improvement fixes.
http://developers.sun.com/prodtech/javatools/jscreator/downloads/hotfixes/index.jsp
Regarding creator on suse 10.1, users have reported that it runs fine:
http://forum.sun.com/jive/thread.jspa?threadID=97305&messageID=334105

Studio Creator installation problem with Fedora 5

Hi,
I tried installing JCreator in Fedora core 5, but I get the problem concerning the graphical installation. I tried installing Jcreator in silent mode using the command line options -is:log log.txt �silent and a sp file which sets the installDir and is.debug to 1 , however the installation does not complete, properly and it does not install the SunAppServer8 directory

I can't recall what they are, but I know that there are some libs that you need to install to get the Creator installation to complete OK on Fedora 5.
I think that I ended up using some libs from either Fedora 3 or 4 since there were no Fedora 5 versions available for the two (and again, I can't recall what they were) libs.
I do know that it is possible though as I have Creator on Fedora 5 right now.

Snap Creator on OVM

Similar Messages

Maybe you are looking for