Description of problem: Version-Release number of selected component (if applicable): How reproducible: Better hadoop performance will result if the rhs-high-throughput tuned-adm profile is set up on each storage node in the cluster Steps to Reproduce: 1. run post Denali version of installer 2. tuned-adm active 3. see the deault profile is active Actual results: Expected results: Additional info:
fixed in 2.15
Version 2.17 of setup_cluster.sh supports a new --profile option. The default is to not change the tuned-adm profile. But, if --profile <profileName> is specified then that profile is verified and set.
I tested it on 2x2, 2x2 with mgmt and brick on the same node and 1x2 with mgmt and brick on the same node. It does not work with the last mentioned layout (the machine with brick and mgmt on the same node stays on default profile), but Im not sure it applies on 1x2 scenarios, but Im writing it here just to be sure. On the other scenarios, it works well (all the nodes are switched to rhs-high-throughput after running ./setup_cluster with --profile option). Please, write, if it apply on 1x2 scenarios also Thank you Jakub Rumanovsky
see also BUG 1150069 COMMENT 11 for clarification about RHS packages removal
Jeff, Could you please provide the doc text in the Doc Text field for this bug?
Jeff, Could you please review the edited doc text for technical accuracy and sign-off?
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHEA-2014-1275.html