[jira] [Commented] (HADOOP-12975) Add jitter to DU's thread
[ https://issues.apache.org/jira/browse/HADOOP-12975?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15241536#comment-15241536 ] Elliott Clark commented on HADOOP-12975: bq. What motivates choosing 50% of the full refresh period as the jitter default? 50% is probably too high, I'll change it to 15%. But a percentage is chosen as it makes the jitter scale with anyone who changes du periods. If it's a set number then someone with a refresh period of days won't get any benefit from the jitter. > Add jitter to DU's thread > - > > Key: HADOOP-12975 > URL: https://issues.apache.org/jira/browse/HADOOP-12975 > Project: Hadoop Common > Issue Type: Sub-task >Reporter: Elliott Clark >Assignee: Elliott Clark > Attachments: HADOOP-12975v0.patch, HADOOP-12975v1.patch > > > Running DU across lots of disks is very expensive and running all of the > processes at the same time creates a noticeable IO spike. We should add some > jitter. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Commented] (HADOOP-12975) Add jitter to DU's thread
[ https://issues.apache.org/jira/browse/HADOOP-12975?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15225299#comment-15225299 ] Colin Patrick McCabe commented on HADOOP-12975: --- Good idea. Can we make the jitter percentage configurable? What motivates choosing 50% of the full refresh period as the jitter default? For people with really long DU periods, this could extend that refresh time by quite a bit. Perhaps jitter should be specified as a maximum number of seconds uniformly randomly added, rather than a percentage of the full refresh period? I think this can get in separately from HADOOP-12973. > Add jitter to DU's thread > - > > Key: HADOOP-12975 > URL: https://issues.apache.org/jira/browse/HADOOP-12975 > Project: Hadoop Common > Issue Type: Sub-task >Reporter: Elliott Clark >Assignee: Elliott Clark > Attachments: HADOOP-12975v0.patch, HADOOP-12975v1.patch > > > Running DU across lots of disks is very expensive and running all of the > processes at the same time creates a noticeable IO spike. We should add some > jitter. -- This message was sent by Atlassian JIRA (v6.3.4#6332)