Re: Travis job time
On Tue, Mar 10, 2020 at 01:05:31PM +0100, Stefano Babic wrote: > Hi Tom, > > On 10.03.20 12:55, Tom Rini wrote: > > On Tue, Mar 10, 2020 at 10:48:43AM +0100, Stefano Babic wrote: > > > >> Hi, > >> > >> was the job time on travis reduced ? Even if jobs are high split, I get > >> several task failures due to "job time". This comprises "non i.MX6 catch > >> all", "catch all ARM" and "catch all AARCH64". Known changes or we > >> should go on and split again and again ? > > > > It's still 50 minutes and I've posted a patch yesterday to address some > > of this (and I'm going to push it today). I also made but didn't test > > yet another split to further reduce things by moving i.MX8 to its own > > job. > > I do not know if we can do a lot of...yesterday I tried several times, > and more jobs failed due to exhausted job time. I just rerun today and > same jobs are successfully. "non-i.MX6 catch all" failed yesterday (> 50 > minutes), it took 15 minutes today to run. It just depends on Travis' > load and there is less we can do it. It is just unsatisfactory and I > agree we should switch to another platform to avoid issues that are > independent from our development. Yeah, that's another part of the problem. Maybe a tutorial on signing up for and setting up Azure to poll an external server would be good. It was rather straight forward I thought when I set things up. -- Tom signature.asc Description: PGP signature
Re: Travis job time
Hi Tom, On 10.03.20 12:55, Tom Rini wrote: > On Tue, Mar 10, 2020 at 10:48:43AM +0100, Stefano Babic wrote: > >> Hi, >> >> was the job time on travis reduced ? Even if jobs are high split, I get >> several task failures due to "job time". This comprises "non i.MX6 catch >> all", "catch all ARM" and "catch all AARCH64". Known changes or we >> should go on and split again and again ? > > It's still 50 minutes and I've posted a patch yesterday to address some > of this (and I'm going to push it today). I also made but didn't test > yet another split to further reduce things by moving i.MX8 to its own > job. > I do not know if we can do a lot of...yesterday I tried several times, and more jobs failed due to exhausted job time. I just rerun today and same jobs are successfully. "non-i.MX6 catch all" failed yesterday (> 50 minutes), it took 15 minutes today to run. It just depends on Travis' load and there is less we can do it. It is just unsatisfactory and I agree we should switch to another platform to avoid issues that are independent from our development. Regards, Stefano -- = DENX Software Engineering GmbH, Managing Director: Wolfgang Denk HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-53 Fax: +49-8142-66989-80 Email: sba...@denx.de =
Re: Travis job time
On Tue, Mar 10, 2020 at 10:48:43AM +0100, Stefano Babic wrote: > Hi, > > was the job time on travis reduced ? Even if jobs are high split, I get > several task failures due to "job time". This comprises "non i.MX6 catch > all", "catch all ARM" and "catch all AARCH64". Known changes or we > should go on and split again and again ? It's still 50 minutes and I've posted a patch yesterday to address some of this (and I'm going to push it today). I also made but didn't test yet another split to further reduce things by moving i.MX8 to its own job. -- Tom signature.asc Description: PGP signature
Re: Travis job time
Hi Stefano, On Tue, Mar 10, 2020 at 5:48 PM Stefano Babic wrote: > > Hi, > > was the job time on travis reduced ? Even if jobs are high split, I get > several task failures due to "job time". This comprises "non i.MX6 catch > all", "catch all ARM" and "catch all AARCH64". Known changes or we > should go on and split again and again ? > Maybe it's time to switch to Azure :) Regards, Bin