Start a Conversation

This post is more than 5 years old

Solved!

Go to Solution

512

August 18th, 2008 09:00

Number of Solutions Enabler deployments.

Hi folks,

Quick question before I open a case with support. Is there a limit to the number of deployments of Solutions Enabler that can be deployed in a SAN eniviornment? The reason I ask is that a number of scripted solutions I run a regular basis are becoming slower and slower to return responses.
In the past year we have deployed a large number of solutions where each host gets a version of Solutions Enabler installed. We now have upto 40 hosts running SE. A lot of them will be called upon once a day to do sync and split of BCV devices (local and remote) alongside masking and so forth. Would this much activity cause performance issues?

Message was edited by:
DavidF

2.8K Posts

August 18th, 2008 12:00

No "known" limits .. however the more syscalls you issue, the longer they'll last .. I have a bright example .. :D
At a customer site we had 90+ hosts issuing clone commands against SRDF/A R2 volumes. Each host manages its own devices and each host issues commands to check health of replica (SRDF status) before actually issuing clone commands ... Customer is scheduling commands (via cronjobs) and when all hosts start sending commands to storage, a single "symdev show" may take up to a minute. From time to time we have to run the same procedure "outside" schedule window (thus when no other host is issuing syscalls against the storage) and the exact same procedure lasts a lot less then before (and symdev returns its output in seconds, as usual) :D

In a word .. it depends .. however the more syscalls, the more they'll last. So the issue isn't with hosts but with syscalls as you noted :-)

15 Posts

August 18th, 2008 10:00

Hi,

We deploy 8 gatekeepers per host as a matter of routine. I personally never agreed with such a solution (>40 hosts with SE installed), and would have preferred a clustered SE environment in each data centre, with seperate installations for TSIMSNAP based solutions. Easier to troubleshoot/manage/secure. But if I am to put my humble case to upper management I need to know if this is having an effect on performance, e.g. is the number of syscalls increasing dramatically as a result etc.

1 Rookie

 • 

20.4K Posts

August 18th, 2008 10:00

i don't think there is a limit, you might have to start adding more gatekeepers for these hosts to spread the load of SE commands. How about consolidating these scripts on just a couple of management hosts because most of the timefinder/masking commands don't have to run on the same server that devices presented to unless you are using some kind of Oracle/Sql/Exchange integrations packages.. Just a thought.

1 Rookie

 • 

20.4K Posts

August 18th, 2008 10:00

hard to say ....it might be that the array is just so busy with regular operations that SE commands take longer to come back ?
No Events found!

Top