"High Load" is when RecoverPoint can't replicate fast enough due to heavy writes. I recommend you investigate and find out what occured during the time the CG went into High Load. What's the available bandwidth between the sites?
A highload is a bottleneck somewhere in the flow, that forces RecoverPoint to drop replication until it clears. it could be on the network pipe (very common) but also on other parts, like the target storage, which is the hardest-working part of the system.
a few questions, as you mentioned SRM:
did the high load occur during an SRM test cycle that was not released for a long time?
did it occur during the booting phase of an SRM test/failover? we found that if the total memory segmented to all VMs exceeds the total ESX machine, it will allocate the VMs swap files on boot, and if that swap is on the replicated LUN (on the target), the IOs will be tracked by RecoverPoint and will cause a highload. VMware has recommended to keep swap files off replication LUNs (on both sites).
etaljic81
1K Posts
0
March 1st, 2012 13:00
"High Load" is when RecoverPoint can't replicate fast enough due to heavy writes. I recommend you investigate and find out what occured during the time the CG went into High Load. What's the available bandwidth between the sites?
EhudKaldor
29 Posts
0
March 1st, 2012 14:00
A highload is a bottleneck somewhere in the flow, that forces RecoverPoint to drop replication until it clears. it could be on the network pipe (very common) but also on other parts, like the target storage, which is the hardest-working part of the system.
a few questions, as you mentioned SRM:
any of these help?