[
https://issues.apache.org/jira/browse/CASSANDRA-7838?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14135111#comment-14135111
]
AMIT KUMAR commented on CASSANDRA-7838:
---------------------------------------
I spent some time looking at these:
* High disk latency -Sigar cannot determine disk io queue lengths on OSX or on
file systems mounted via NFS on linux boxes it returns -1, ditto for disk
reads, writes.
* Swap enabled -we can determine that with accuracy on Unix/Linux and OSX
* Ulimit -open files/max user process, those I could get on linux and OSX
What ulimit values do we think are sane?
Can one of you give disk latency a try on EBS disks? I can commit the changes
to a branch and we can figure out what would be most useful.
Thanks
Amit
> Warn user when disks are network/ebs mounted
> --------------------------------------------
>
> Key: CASSANDRA-7838
> URL: https://issues.apache.org/jira/browse/CASSANDRA-7838
> Project: Cassandra
> Issue Type: Improvement
> Reporter: T Jake Luciani
> Priority: Minor
> Labels: lhf
> Fix For: 3.0
>
>
> The Sigar project let's you probe os/cpu/filesystems across the major
> platforms.
> https://github.com/hyperic/sigar
> It would be nice on start-up to use this to warn users if they are running
> with settings that will make them sad, like Network drive or EBS on Ec2.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
