I run /opt/drbl/sbin/dcs succesfully with default settings and it found image file. Then I chose network boot from client and when Partclone started it freezed.
Just text "Starting to restore image (-) to device (/dev/sda2)" appeared and the cursow stayed to blinkin below.
What did go wrong?
Partclone 0.2.49 is old. Did you try the latest Clonezilla live ? E.g. 2.1.0-26, and it comes with Partclone 0.2.58.
Hello, I also have the same problem. My partclone is 0.2.59.
Any advice ?
Unicast is OK. Only Multicast has the freeze situation. I had tried several switches. I also try both server and client are VMs. But no help.
Thanks for any suggestions.
So you meant this issue is reproducible in your VM?
Besides, did you try Partclone 0.60?
I'm using the latest alternative Ubuntu version of Clonezilla-live 20140415-saucy.
It is a year since the thread open but I'm still seeing the same symptom. I view this issue as not caused by Partclone. The client HD is being initialized. The last message on the PartClone screen indicates the hang is at the start of cloning.
The issue happens only when using multicast as noted above. Unicast is working for this image.
I'm going to consider the PartClone version and try using the latest stable Clonezilla-live version 2.2.2-37
The issue was not caused by using the latest alternative Ubuntu version of Clonezilla-live 20140415-saucy. I had forgotten to run "drbl-ocs" to start the multicast process.
root 23026 0.0 0.0 1000 364 pts/1 S 14:31 0:00 udp-sender --full-duplex --min-clients 5 --max-wait 300 --interface eth1 --nokbd --mcast-all-addr 22.214.171.124 --portbase 2232 --ttl 1
drbl-ocs -b -g auto -e1 auto -e2 -x -r -j2 -p reboot --clients-to-wait 5 --max-time-to-wait 300 -l en_US.UTF-8 startdisk multicast_restore Image-04-04-13-img sda
I am getting significantly slower NFS performance using mulicast vs. unicast.
"I am getting significantly slower NFS performance using mulicast vs. unicast." -> What did you mean? Could you please describe more?
All of the PCs are identical hardware. I am also starting to deploy laptops as well. I'm typically working with with 10-14 Gb compressed Windows images.
During a single unicast restore I am seeing NFS transfer rates of 6+ Gb/min. I've even done 3 different images restoring while a 4th PC is creating an image with 5+ Gb/min throughput which restore from PXE boot in less than 4 minutes each.
With a multicast_restore a single PC will get 1.5 Gb/min with no other activity on the private network. I've only tested multicast_restore with a maximum of 5 PCs and the transfer rate drops to as low as 1.25 Gb/min. Using multicast takes 22-30 minutes to complete. I will be able to restore up to 23 PCs soon as this is a new DRBL server and the lab is still being set up.
Also, after completing a multicast_restore the udp-sender process on the DRBL server is terminated. I'm searching for a solution to have it restart upon demand from a new client. I don't want to give operators access to the DRBL server or kludge a way to have a udp-sender process available. Let me know if you would like me to open a separate thread for this.
"searching for a solution to have it restart upon demand from a new client. " -> In "dcs", you can enter expert mode, and choose the option "-y1" when starting clonezilla-start mode.
BTW, there are some factors could affect multicast efficiency. One of them is the performance of network switch, another one is on the parameters of udpcast. You can tune that in /etc/drbl/drbl-ocs.conf, especially the parameters "udp_sender_extra_opt_default".