DataDomain and High Network Utilization mystery

Question

DataDomain DD7200 DDOS 5.4.4.3 Networker 8238 on Windows 2012r2 Hi all, Our Networking team recently came to me and stated that my DataDomains were chewing up more bandwidth than anything else, possibly causing a bottleneck on our 10gig link between our two datacenters. I can also see this through our Solarwinds monitoring tool. The odd thing is that they are seeing this regularly on Monday afternoons around 4:30. I have a few mtree replications set up that run regularly, so that would not be it. I run my full Networker backups on Fridays and Saturdays, but I dont run any cloning jobs on Monday afternoons, other than daily vm cloning. Our Oracle DBA's do also back up their databases to my datadomains using Rman. Some of their databases are 30 to 40tb is size, but this has always pretty much been the case, so nothing really changed there. Plus, they tell me that they dont perform any of their own cloning on Monday afternoons, and that their cloning is done through the mtrees that I created for them. In short, is there a way that I can tell what it is that one DataDomain is passing to the other, that is eating up bandwidth, only on Monday afternoons at 4:30?? Thanks!

Ryan_Johnson · Answer

I would start with logging into the data domain and running commands to view the replication history.

Command "replication show history all" will show mtree, directory and collection replication context history. You have an options of showing individual replication context. By default it will show the last day. You could add "duration 168" to show the last week.

For Networker clone jobs (data domain managed file replication) you can run command "ddboost file-replication show history".

Toddman214 · Answer

Thanks Ryan, I cant seem to get that syntax right.

sysadmin@ddc00ddma904# replication show history duration 24

**** Missing one or more required arguments.

Incomplete syntax "replication show history duration 24", showing help:

replication show history { | all}

[duration {hr | min}] [interval {hr | min}]

Show replication performance history

Ive looked up and tried various combinations, but still leaving something out. Any idea?

Toddman214 · Answer

Awesome, thanks for the clarification, repl-man!

Dell-JoeB · Answer

Also I noticed you are running DDOS 5.4.4.3 keep in mind that it will drop out of support at the end of June.  Also 5.5x goes out of support at the end of December.

Dell-JoeB · Answer

The command you are looking for is:

#replication show history all duration 24hr interval 24hr

This is also in the daily autosupports. This will only account for mtree, directory and collection replication. If you have clone control jobs you will need to look at the ddboost section of the asups to see what you are using there. I have also seen Solarwinds not pull accurate data on bandwidth utilization when pulling utilization data. There are alot of forums out there that cover this issue and how to correct that.

Data Domain

Was this post helpful?