Veeam Replication Stays @ 0% for 40 Minutes.

[cag]

I am testing Veeam Backup and Replication using 1 Production Server and 1 Test DR Server. I have the following problems. Speed is so slow.
Step 1 Start the job, Sits @ 0% for 25- 30 Minutes. I am not sure what is happening during this 25 Minutes but the CPU is at almost 2-12 % I dont notice much disk activity but network activity is about 400-500KB/S solid.

I have tried every mode from Direct SAN / VM Appliance / Netowrk. Local / LAN / WAN Targets, but I always get 0% for a good 45 Minutes. The VM is a first time replication of 32GB Windows XP VM.

I have tried the VM Agent at the source location and the DR Location with the same results. Seems to do nothing for 45 Minutes.

Network activity at the DR Server is up and down and MAX at 250KB / S for a few seconds and then nothing.

Source Server is 2x QUAD 2.4 Xeons with a direct connect SAN 24GB memory and the Target is a 2xDual 3GZ Xeons with GB RAM and no VM's Running.

Veeam Server is running as a VM on another server idential to the Source server connected to the same SAN.

[Vitaliy S.]

Hello Conrad,

I have tried every mode from Direct SAN / VM Appliance / Netowrk. Local / LAN / WAN Targets, but I always get 0% for a good 45 Minutes.

Well...this definitely indicates that something is wrong with communication between your backup server and vCenter Server. Usually a "preparation" baclup period takes no more than a couple of minutes (in worst case scenario).

Could you tell me whether you see the same behaviour for every VM being backed up? I would strongly recommend reaching out our support team and provide them with the replication job log for further investigation.

Thanks.

[cag]

I have sent the logs, but without phone support (email support sucks) and no remote screen sharing so they can see what I see, the process is painful. Seems the application has no useful tools to diag. The logs are useless to a human. I can complete the whole replication process manually faster than the product.

eg. I take a snapshot of the VM. Open my datastore browser and copy the VM files. Once completed I delete the snapshop.

here is an example of what I see.

3 of 4 files processed

Total VM size: 90.00 GB
Processed size: 9.53 KB
Processing rate: 0 KB/s
Backup mode: HOTADD/NBD with changed block tracking
Start time: 3/26/2011 9:03:48 AM
Time remaining: 19:58:22

Replicating object '[SATA7200] HMI SBS 2010/HMI SBS 2010.vmdk'

This is after 55 minutes of the job running. process 9.53KB

So where is the content going ??????

[Gostev]

Hello, please note that this is not a support forum. Please, do not post log snippets as requested when you click New Topic. For assistance with any technical issues requiring debug logs investigation, please work with our technical support directly.

We do of course provide phone support (for phone numbers, please refer to Support page on our web-site). Our support engineers also do leverage webex whenever needed if customer policy allows for this, because this is more convenient indeed. However, please note that for Standard Support program, support is only provided during business days nad business hours as defined in the support policy.

Concerning the above, I would assume the issue is a link speed. Statistics is only updated every 1 percent of VMDK file. So, if you effective link speed to the DR location is just a few Mb, no statistics updates for extended periods of time are to be expected. Besides, first update will be delayed because of extra operations before that (replica initialization, guest freeze, VM snapshot, etc). However, if you link speed is fast, then the reason might different. Our technical staff will be able to tell exactly what is happening from the full log files.

Thanks.

[cag]

Updates every 1% is a dumb design. What Human would expect that kind of response. Most apps update based on time intervals. EG VMWARE Client Performance Screens. Windows Performace Screens, Linux etc / etc.

The interface needs a new design. Should update at time intervals and display # KB / MB transfered of the Total Amount to transfer, and calculate the % complete.

This way I know how much data must move and I can see if a process has stoped based on KB transfered. When the amount stops incrementing then I know I have a problem.

This would also allow me to monitor total data that must move each day to help plan bandwidth requirements to support the amount of data that changes, but also to know that changed block backup / replication is working. I have run some test where the second replication takes as long as the first replication on a static VM.

[Gostev]

Already answered here. To be fair though, all product you are listing are monitoring products. Of course, Veeam Monitor (for example) provides realtime updates as well. But for BC/DR solution, real-time performance monitoring is by far not the key feature backup jobs is not something you need to be monitoring in realtime everyday - they are set and forget thing... summary stats in job report is all what matters.