Staggering the repairs also gives the DynamicSnitch a chance to route around nodes which maybe running slow.
Cheers ----------------- Aaron Morton Freelance Developer @aaronmorton http://www.thelastpickle.com On 29/08/2012, at 11:19 AM, Omid Aladini <omidalad...@gmail.com> wrote: >>> Secondly, what's the need for sleep 120? >> >> just give the cluster a chance to settle down between repairs... >> there's no real need for it, just is there "because". > > Actually, repair could cause unreplicated data to be streamed and new > sstables to be created. New sstables could cause pending compactions > and increase the potential number of sstables a row could be spread > across. Therefore you might need more disk seeks to read a row and > have slower read response time. If the read response time is critical, > it's a good idea to wait for pending compactions to settle before > repairing other neighbouring ranges that overlap replicas. > > -- Omid > >> -- >> Aaron Turner >> http://synfin.net/ Twitter: @synfinatic >> http://tcpreplay.synfin.net/ - Pcap editing and replay tools for Unix & >> Windows >> Those who would give up essential Liberty, to purchase a little temporary >> Safety, deserve neither Liberty nor Safety. >> -- Benjamin Franklin >> "carpe diem quam minimum credula postero"
