That modification may make sense, but it still seems that the calculation of the existence of a live server is incorrect - if there is a live region server and a dead region server, the net result isn't "no region server".
-- Jon On Jul 12, 2014, at 6:42 AM, Ted Yu <[email protected]> wrote: > Alternatively the agent detects whether previous port is available, reuses > the same port if it is. Otherwise fallback to current behavior. > > This would work in single tenant case. > > Cheers > > On Jul 12, 2014, at 3:11 AM, Steve Loughran <[email protected]> wrote: > >> ...maybe the agent could be set up to perform a sleep for a while if a port >> is in use, in the hope it will be cleaned up. >> >> >> On 12 July 2014 06:34, Sumit Mohanty <[email protected]> wrote: >> >>> It cannot be guaranteed that the previous port will still be available when >>> the application is thawed - in reality the application can be thawed in few >>> minutes or even few days. So I think reusing the old port might be a risk. >>> >>> Isn't this a case for the QE script to change? Inherently, for yarn apps, >>> they need to handle dynamic host/port. >>> >>> >>> On Fri, Jul 11, 2014 at 9:16 PM, Ted Yu <[email protected]> wrote: >>> >>>> Hi, >>>> When I was debugging existing QE hbase test script, I found that in the >>>> following situation (create - freeze - thaw) it was determined that there >>>> was no live server: >>>> region server, hor10n03.gq1.ygridcore.net,42175,1405108984098, was >>>> considered dead by the new master due to 'freeze' action >>>> the new region server, hor10n03.gq1.ygridcore.net,46329,1405120269524 >>>> <http://hor10n03.gq1.ygridcore.net:60941/>, was live however master >>>> didn't remove the first one from the dead servers list due to port not >>>> matching. >>>> QE script drew the conclusion because 1(live)-1(dead) = 0 >>>> >>>> You can observe this scenario here: >>>> http://hor10n01.gq1.ygridcore.net:50938/master-status >>>> >>>> Since region server was brought up on the same node and the previous port >>>> was still free: >>>> >>>> [hortonzy@hor10n03 ~]$ sudo netstat -tulpn | grep 42175 >>>> [hortonzy@hor10n03 ~]$ >>>> >>>> I think proper action should be to reuse the previous port when thawing. >>>> >>>> Please comment. >>> >>> -- >>> CONFIDENTIALITY NOTICE >>> NOTICE: This message is intended for the use of the individual or entity to >>> which it is addressed and may contain information that is confidential, >>> privileged and exempt from disclosure under applicable law. If the reader >>> of this message is not the intended recipient, you are hereby notified that >>> any printing, copying, dissemination, distribution, disclosure or >>> forwarding of this communication is strictly prohibited. If you have >>> received this communication in error, please contact the sender immediately >>> and delete it from your system. Thank You. >> >> -- >> CONFIDENTIALITY NOTICE >> NOTICE: This message is intended for the use of the individual or entity to >> which it is addressed and may contain information that is confidential, >> privileged and exempt from disclosure under applicable law. If the reader >> of this message is not the intended recipient, you are hereby notified that >> any printing, copying, dissemination, distribution, disclosure or >> forwarding of this communication is strictly prohibited. If you have >> received this communication in error, please contact the sender immediately >> and delete it from your system. Thank You. -- CONFIDENTIALITY NOTICE NOTICE: This message is intended for the use of the individual or entity to which it is addressed and may contain information that is confidential, privileged and exempt from disclosure under applicable law. If the reader of this message is not the intended recipient, you are hereby notified that any printing, copying, dissemination, distribution, disclosure or forwarding of this communication is strictly prohibited. If you have received this communication in error, please contact the sender immediately and delete it from your system. Thank You.
