Обсуждение: buildfarm stats
Sometime in late June the buildfarm passed 50,000 builds reported on. Here are stats over the lifetime. cheers andrew ym | builds | reporting_members --------+--------+-------------------2004-10 | 181 | 62004-11 | 1533 | 122004-12 | 2468 | 272005-01 | 2432 | 302005-02 | 1367 | 252005-03 | 1970 | 282005-04 | 2297 | 282005-05 | 2348 | 282005-06 | 2543 | 342005-07| 3038 | 432005-08 | 2748 | 412005-09 | 2104 | 362005-10 | 2597 | 362005-11 | 2214 | 352005-12 | 2534 | 382006-01 | 3338 | 412006-02 | 2843 | 402006-03 | 2536 | 432006-04 | 2978 | 452006-05| 2816 | 452006-06 | 3338 | 50
> -----Original Message----- > From: pgsql-hackers-owner@postgresql.org > [mailto:pgsql-hackers-owner@postgresql.org] On Behalf Of > Andrew Dunstan > Sent: 03 July 2006 23:56 > To: PostgreSQL-development > Subject: [HACKERS] buildfarm stats > > > Sometime in late June the buildfarm passed 50,000 builds reported on. > Here are stats over the lifetime. Thanks for the stats Andrew. Out of interest, can you easily tabulate the number of failures against OS? Regards, Dave.
> Thanks for the stats Andrew. Out of interest, can you easily tabulate > the number of failures against OS? Or, more generally, even put a dump of the DB (without personal infos of course :) somewhere? Bye, Chris. PS: and don't say you're running it in MySQL ;)
On Tuesday 04 July 2006 22:14, Chris Mair wrote: > > Thanks for the stats Andrew. Out of interest, can you easily tabulate > > the number of failures against OS? > > Or, more generally, even put a dump of the DB (without personal infos > of course :) somewhere? > > Bye, Chris. > > PS: and don't say you're running it in MySQL ;) Well as the host, I guarantee you that it is NOT running mySQL :) Sincerely, Joshua D. Drake > > > > > ---------------------------(end of broadcast)--------------------------- > TIP 2: Don't 'kill -9' the postmaster
Joshua D. Drake wrote: >On Tuesday 04 July 2006 22:14, Chris Mair wrote: > > >>>Thanks for the stats Andrew. Out of interest, can you easily tabulate >>>the number of failures against OS? >>> >>> >>Or, more generally, even put a dump of the DB (without personal infos >>of course :) somewhere? >> >>Bye, Chris. >> >>PS: and don't say you're running it in MySQL ;) >> >> > >Well as the host, I guarantee you that it is NOT running mySQL :) > > > > but it is about 2Gb of data, so just putting a dump cleaned of personal data somewhere isn't really an option. I could arrange a dump without the diagnostics, in these 2 tables: system: < name | operating_system | os_version | compiler | compiler_version | architecture > build: < name | snapshot | stage | branch | build_flags > (stage in the latter table is OK on success or the name of the stage that failed otherwise). But what do you want it for? And do you want it one-off or continuously? cheers andrew
> but it is about 2Gb of data, so just putting a dump cleaned of personal > data somewhere isn't really an option. > > I could arrange a dump without the diagnostics, in these 2 tables: > > system: < name | operating_system | os_version | compiler | > compiler_version | architecture > > build: < name | snapshot | stage | branch | build_flags > > > (stage in the latter table is OK on success or the name of the stage > that failed otherwise). > > > But what do you want it for? And do you want it one-off or continuously? Nothing important at all. I'd just thought about a few interesting stats, like failures vs OS (as the first poster said) or failures vs gcc version or timings vs. arch / RAM or gcc version, etc. For the timings I guess there are some timestamps embedded that might be extracted... But I didn't really think about it, before posting (classic mailing list syndrome ;) Bye, Chris.
Chris Mair wrote: >>but it is about 2Gb of data, so just putting a dump cleaned of personal >>data somewhere isn't really an option. >> >>I could arrange a dump without the diagnostics, in these 2 tables: >> >>system: < name | operating_system | os_version | compiler | >>compiler_version | architecture > >>build: < name | snapshot | stage | branch | build_flags > >> >>(stage in the latter table is OK on success or the name of the stage >>that failed otherwise). >> >> >>But what do you want it for? And do you want it one-off or continuously? >> >> > >Nothing important at all. >I'd just thought about a few interesting stats, like failures vs OS >(as the first poster said) or failures vs gcc version or timings vs. >arch / RAM or gcc version, etc. For the timings I guess there are >some timestamps embedded that might be extracted... > >But I didn't really think about it, before posting (classic >mailing list syndrome ;) > > > > We don't have any timing info. For now this goes on the wishlist. cheers andrew