Replication Job Wont Stop!

[cag]

How do you stop a repliaction job. I tried the right mouse click stop. But nothing happens, I have waited more than 30 minutes and still nothing. All I can do is shutdown the server and restart. After the restart the next replication job fails with a varity of differnet problems. Cannot delete files or files are missing. Forced to delete the VM and start again. This is not very productive for a DR Site.
Thankfully this is only a test setup from a production system of 32GB, but if this is how stable Veeam is how can anyone use this when a server could have 200-1TB of data.

[mplep]

I have had to wait quite a while for a replication job to stop on occasions, sometimes a couple of hours.

[Vitaliy S.]

Basically, if you face any technical issues with either stopping or continuing the backup/replication job you should contact our technical team for assistance. There is a hardway to stop the job (end all veeamagent processes), but it is not a recommended way.

[cag]

Seems there is no good way of stoping any job. This product is not ready for Production Use. Seems no testing has been done for failure recovery. EG Network Link or server crashes during a job. Since Job can take hours to complete it is reasonable to assume the product will crash, but please recovery from where it left off. Dont make me delete my replica and start over. I except this from a $50.00 program not one that costs $thousands. I have contacted you technical team they told me to delete all the replica files on the target and try again. ( That is not support )

[Gostev]

Hi.

Of course, the product does supports replica recovery after server crash, job crash or link down condition - since version 1. The engine was built with this in mind and is designed to be able to handle these (definitely quite common) issues. However, under certain circumstances there could be exceptional cases occasionally when full resync is required. We are addressing the full resync requirement in the next release completely.

If you really need to stop the job (which you should not need to do under normal circumstance), then depending on link speed, replica size and replica target (ESX or ESXi) you will have to wait in some cases because agents process stop commands only once every 1% of VMDK size, this is by design. Last thing you want to do is reboot your server, this will not help because it does not shutdown the target agent. Agent will remain running for quite some waiting for source agent to respond (which is needed to support replication over very slow link) and keep the files locked.

If you are not happy with support you are getting from your support engineer, you can always request a callback from support manager.

Thanks.

[cag]

Stops every 1% seems like a bad design. What normal human would except that kind of response. When I click stop I except now. Every other backup app I have ever used stops within a minute. Even VDR from VMWARE stops within a resonable time.

But based on your information I now understand why it takes my replication job 30 - 40 minutes before it seems to start processing. Well its your interface design. instead of doing like every other app in the world updated at time intervals your screen updates are based on 1% completion intervals. Veeam's Approach requires explanation when it should be standard like every app that has a process bar. The interface needs to display x KB/MB of Total Size and calculates the % done. This way I know the process is not hung because I can see the data amount transfered is increasing.

If your inferace doesn't change then you need to display a message when the process will stop. eg x minutes based on how long %1 takes to process. As data stores hit the TB sizes and the amount of bandwidth a backup or restore process could take it seems stupid to have to wait the 1% increment. this could be 30-60 minutes over wan connections, Meantime your network bandwidth may impact normal operations.

[Gostev]

I actually hate 1% updates myself, so I do not disagree with you. I've been pushing devs to resolve this, but it was hard to justify touching whole grail (processing engine) over UI improvement. Anyway, this feature is on the drawing board to come along with future engine updates.