P4500 snapshots getting only 4kb/s transfer

[Chris Bullock_1]

Setup
1 set of P4500 LeftHand boxes living on my Gigabit lan in 2 different buildings.
One is my primary and one is secondary. I am snapshotting from primary to secondary. I am only getting about 8kb/s transfer speed. I have gone in and edited bandwidth until I have hit "max" in the management group.
I haven't even put any data on these boxes yet and it's taking over 20 minutes to copy 1.25mb of data.

[René Loser]

Hello Chris,

How does your network looks like?
You may have a drawing?

regards,
reNe

HP PreSales Switzerland

[Olvi_1]

If you have edited the bandwidth to "max" I guess you have edited the local bandwidth level. That does nothing when you make remote snapshots.

Check what is your Remote Management Group Bandwidth Level (look attachment). You can edit it to whatever you like it, calculate to optimum or use default speeds depending on your bandwidth

Now that you haw gigabit LAN between buildings is there a good reason why do you want to make config this way? With one node per cluster gives you no fault tolerance. Consider making Multi-site cluster (inside one Management Group) + install Failover Manager and use network RAID. This way one node can go offline without interrupting IO to hosts.

[Chris Bullock_1]

I have remote bandwidth set to Gigabit ethernet. My switches are running clean and there appears to be no problems on the switch port level.

[Olvi_1]

Hi!

I noticed you had network problems earlier. Is this issue solved and what was the problem?
http://forums13.itrc.hp.com/service/forums/questionanswer.do?threadId=1411547

Here is some hints for troubleshooting the problem.
- Ping node to a another (gig LAN, time should be next to nothing)
- Try different NIC bondings (ALB recommended)
- Check alert logs on nodes
- Make LUN to Sec. MG and Remote Snapshot to Pri. (any difference)
- Make Remote Snapshot inside Management groups (check for speed rates)
- Bring node from secondary site to primary and connect to same switch(es) (to eliminate possible network problems)
- If possible transfer data between sites on some other devices such as desktops. Any problem there.

If non of those steps give a clue and if there is no performance issues at primary site/node I would start to suspect faulty node.

[Chris Bullock_1]

What we have noticed this morning is that if we let the system create the remote volume this fixes our problem. We are now getting anywhere from 100-200mb/sec transfers. We originally manually created the Luns and volumes on the remote system as thin provisioned and then mapped to that for the remote copy.

[Frank Verstraten]

I Am having the same issue with P4500 G2 nodes.
After setteng the bandwith to max values all I get is 300Kbps. I can't get the max speed of 1 Gbps.

[Bryan McMullan]

Frank, what is your network layout? Are you sure you don't have any network issues between the management groups (Duplex mismatches cause by the link provider have sometimes caused issues like that for me in the past)?

[Frank Verstraten]

I used different setups. Even if I install all iscsi cables in one switch I still have the same results.
The config is now 2 nodes using netraid 2 in one management group and one remote node in another man group to do the remote copy.
All the nodes are in the same vlan.
The switches used are procurves 2910.The bond settings are ALB on all the LH nodes. SanIQ is 8.5.And the duplex settings are set to auto.

[Bryan McMullan]

@Frank -

Seems to be a number of issues lately with remote copy. If you have everything connected to one switch and still have an issue with only 300kbps....either the switch is misconfigured somehow or you should open a case with HP/LH support. That is terribly low and indicates that there is something very wrong.