- Posts: 3
- Liked: 1 time
- Joined: Jun 11, 2018 5:59 am
- Full Name: Andreas Buetler
We are on a translation from a HPE StoreOnce to an ExaGrid Backup Device.
For the retention we have a backup copy job from ExaGrid to the StoreOnce. In the new backup concept we have also changed the swtich that every traffic is separate from the production envoirement.
The backup proxy is connect over 2 10GBit ethernet link configured in dynamic LACP.
In the default config we had very bad copy performance. The jobs was running with 1 to 4 MB/s. Now we have found that the default load-sharing policy on the HPE Comware 5 based switch was the problem.
Default was: packet type-based sharing
Now i have changed the criteria to source-ip address and destination-ip address. With this it works fine.
My question have someone a recomendation for LACP load-sharing polcy configuration on HPE switches that works great with VEEAM?
Veeam case id was 03144394
- Service Provider
- Posts: 338
- Liked: 67 times
- Joined: Jan 31, 2011 11:17 am
- Full Name: Max
Regardless of your load balancing algorithm your connection will always go over a single interface. If you have multiple connections in the worst case those could all be put on the same interface, while having the other interface in idle; in other words you can't evaluate load. With Windows Server 2012 and above I would go with switch independent dynamic teaming.
Users browsing this forum: No registered users and 19 guests