VMware Datarecovery Kernel Panic possible solution

Since May ( Post ) I was facing multiple Kernel panic’s in my VDR appliance. Now I got something like a fix which is running quite smooth in my environment.

Steps:
1. Deploy a new VDR 2.0 appliance
2. SSH in or log into the console (root/vmw@re) 3. Install this kernel.
rpm -i http://elrepo.org/linux/kernel/el5/x86_64/RPMS/kernel-ml-2.6.39-4.1.el5.elrepo.x86_64.rpm
4. Set the kernel as the default one. (You can manually select this one at. The command just sets grub to boot the first kernel in the list) sed -i ‘s/default=1/default=0/’ /boot/grub/menu.lst 5. Reboot.
6. Confirm we are running on the new kernel # uname -r 2.6.39-4.1.el5.elrepo

Give it a try and if possible provide me with your experiences.

VMware DataRecovery Kernel Panics

I’m currently facing an problem with my vdr that it runs into a kernel panic.

It looks like an issue with the hot add function which add’s vmdks from virtual machines.

The response from vmware was to set to parameters in the datarecovery.ini file.

Location: /var/vmware/datarecovery/datarecovery.ini

If the file doesn’t exist you can create it.

Parameters to be added to this file:

SerializeHotadd=1
DisableHotaddCopy=1

After the file is created and the parameters have been inserted the datarecovery service has to be restarted.

service datarecovery stop
service datarecovery start

I’m currently testing these settings. Keep you in the loop. :-)

long time no post … :-(

I’m very busy at work as we are currently in major reorganisation. Personal for myself everything has changed. I’m now in a new department including a new boss and new responsibilities. But now everything settles and I have time for new stuff like my preparation for the VCP 5 exam. :-)

As preparation I have upgraded my Homelab completely to vSphere 5. Both Microserver and the ML115 G5 are running ESXi 5 with the latest patches. The vCenter is running as virtual machine located on the ML115 G5 as this box has much more power than the Microservers. Next step is setup of a shared storage as this is still missing in my homelab and the VDR for backups.

Hopefully I’m able to finalize these things asap.

Memory leak in VDR 2.0

Yes there is a memory leak in the current Version of VDR 2.0. After some days the VDR crash’s with the Message “Out of Memory”.

Get PID of datarecovery process

ps -ef | grep datarecovery

Check with pmap to display information about the address space of a process

pmap -x PID

Search for “anon” messages. Anon messages are blocked Memory spaces.

Current workaround is to schedule an maintenace of the destinations and then restart the VDR Appliance on a daily base. It should be possible to use scheduled tasks for the automated restart.