Hyper-V Replica - Understanding Performance

i running 2 servers microsoft hyper-v server 2012 (core) , using hyper-v replica between 2 servers.  both ends have 30/5 mbps connections (down/up).

there 1 vm on each server being replicated.  most of changes/data accumulated on servera.

i setup servera , serverb side side first replication , moved serverb remote location.

recently there lot of data added servera (around 300-400 gb).  over past several weeks more data has been added.  in checking replica stats showed peak of around 980 gb.

around september 9th replication stopped while serverb showed vm "modifying resource."  this continued several hours.

afterwards replication resumed.  (i don't recall if had manually restart replication or if resumed on own.)  but still had several hundred gb replicate.

last night replication around 95%.  a few hours later assume must have hit 100% when checked found replication had stopped , serverb again showing servera vm "modifying resource."

i thought should have meant full 980 gb of data had been replicated.  when replication resumed if "snapshot" had completed.  but in checking stats still saw following.

size of data yet replicated: 616.64 gb
last synchronized at: 09/09/2013 10:22:01 pm

so after running since @ least september 8th amount of replication completed 1/3 of amount of data.  and date moved 6 days.

i haven't added data servera recently, anticipate in next week or 2 will.  at rate i'm wondering if serverb ever catch , servera vm replicated?

when replication resumed last night showed 2% completed.  after running overnight (about 6 hours) shows 4% completed.

when making estimates performance thought 5 mbps upload speed should work relatively small amount of data need replicate.  but i'm not sure going adequate keep up?

especially after seeing once replication reaches 100% doesn't seem mean pending data see going have been sent.  only "snapshot" have completed?

has else done similar testing , have got insight can share these topics?

---

theokrat

hi theokrat

on sept 9th when replication stopped, there event viewer message on primary server, under applications , services logs->microsoft->windows->hyper-v-vmms->admin. depending on initial size of vm, there chance vm might have gone resync required state. might find eventviewer message against vm.

depending on first event viewer message, can debug happened after event. vms typically configured resync during non-business hours. maybe resync started then. based on response , vm size, can explain % numbers.

praveen

Windows Server  >  Hyper-V