Thread: upcoming postgresql.org infrastructure migration
Hi all, The sysadmin team is about to start a fairly large scale migration of the following public facing services to a new datacenter: yum.postgresql.org - running the yum repository information website xindi.postgresql.org - providing wiki.postgresql.org gemulon.postgresql.org - providing git master server norkan.postgresql.org - providing ns2.postgresql.org selek.postgresql.org - website search infrastructure omicron.postgresql.org - providing the PgUS / PostgresOpen website procyon.postgresql.org - developer.pgadmin.org babel.postgresql.org - Translation team website faynos.postgresql.org - Download Mirror arkaria.postgresql.org - running the Mailinglist archives backend rakhar.postgresql.org - providing planet.postgresql.org paxsor.postgresql.org - providing www.pgadmin.org, ftp.pgadmin.org coridan.postgresql.org - running the commitfest app devos.postgresql.org - providing Personal homepages for developers corvan.postgresql.org - Coverity build and submission system brentalia.postgresql.org - running the postgresql.org buildfarm website rubicun.postgresql.org - redmine.postgresql.org jerido.postgresql.org - providing jdbc.postgresql.org makus.postgresql.org - .us inbound mail relay for postgresql.org related domains zakdorn.postgresql.org - Frontend for main website atalia.postgresql.org - staging and testing server for apt.postgresql.org The migration is required because the long term infrastructure sponsor where we host those systems is shutting down the datacenter the current servers are in. We therefore need to migrate the systems to a new datacenter of theirs, but we will do everything we can do to minimized downtime for non-redundant services, but due to the large number of public and internal systems we will not be able to inform about every single migration. If you however note a particular service to be unavailable or impacted for an extended period of time - please dont hesitate to tell us! The migration is starting immediately and is expected to be finished by the end of next week. Stefan
Hi, On 2018-10-16 22:04:39 +0200, Stefan Kaltenbrunner wrote: > gemulon.postgresql.org - providing git master server Could you give a little warning before doing so? It'd be annoying to push a backpatched patch to only half the branches, to push half a patch series, or to push something somewhat experimental that'll turn half the buildfarm red without being able to fix it in time... Greetings, Andres Freund
On 10/16/18 11:40 PM, Andres Freund wrote: > Hi, > > On 2018-10-16 22:04:39 +0200, Stefan Kaltenbrunner wrote: >> gemulon.postgresql.org - providing git master server > > Could you give a little warning before doing so? It'd be annoying to > push a backpatched patch to only half the branches, to push half a patch > series, or to push something somewhat experimental that'll turn half the > buildfarm red without being able to fix it in time... yeah - for gitmaster we will make a seperate announcement... Stefan
On 10/16/18 10:04 PM, Stefan Kaltenbrunner wrote: > Hi all, > > The sysadmin team is about to start a fairly large scale migration of > the following public facing services to a new datacenter: > > > yum.postgresql.org - running the yum repository information website > xindi.postgresql.org - providing wiki.postgresql.org > gemulon.postgresql.org - providing git master server > norkan.postgresql.org - providing ns2.postgresql.org > selek.postgresql.org - website search infrastructure > omicron.postgresql.org - providing the PgUS / PostgresOpen website > procyon.postgresql.org - developer.pgadmin.org > babel.postgresql.org - Translation team website > faynos.postgresql.org - Download Mirror > arkaria.postgresql.org - running the Mailinglist archives backend > rakhar.postgresql.org - providing planet.postgresql.org > paxsor.postgresql.org - providing www.pgadmin.org, ftp.pgadmin.org > coridan.postgresql.org - running the commitfest app > devos.postgresql.org - providing Personal homepages for developers > corvan.postgresql.org - Coverity build and submission system > brentalia.postgresql.org - running the postgresql.org buildfarm website > rubicun.postgresql.org - redmine.postgresql.org > jerido.postgresql.org - providing jdbc.postgresql.org > makus.postgresql.org - .us inbound mail relay for postgresql.org related > domains > zakdorn.postgresql.org - Frontend for main website > atalia.postgresql.org - staging and testing server for apt.postgresql.org > > > The migration is required because the long term infrastructure sponsor > where we host those systems is shutting down the datacenter the current > servers are in. > We therefore need to migrate the systems to a new datacenter of theirs, > but we will do everything we can do to minimized downtime for > non-redundant services, but due to the large number of public and > internal systems we will not be able to inform about every single > migration. If you however note a particular service to be unavailable or > impacted for an extended period of time - please dont hesitate to tell us! > The migration is starting immediately and is expected to be finished by > the end of next week. this migration has been completed for all systems - and there is no currently known remaining service impact. If somebody notices something that stopped working or is behaving differently dont hesitate to tell us ;) Stefan
Stefan Kaltenbrunner <stefan@kaltenbrunner.cc> writes:
> this migration has been completed for all systems - and there is no
> currently known remaining service impact. If somebody notices something
> that stopped working or is behaving differently dont hesitate to tell us ;)
Just nosing around in stuff I use ... it seems like praxis' image
of the buildfarm database is not up to date. The latest build
snapshot it has is about 10 hours old:
pgbfprod=> select max(snapshot) from build_status_log;
max
---------------------
2018-10-20 09:08:56
(1 row)
regards, tom lane
Greetings, * Tom Lane (tgl@sss.pgh.pa.us) wrote: > Stefan Kaltenbrunner <stefan@kaltenbrunner.cc> writes: > > this migration has been completed for all systems - and there is no > > currently known remaining service impact. If somebody notices something > > that stopped working or is behaving differently dont hesitate to tell us ;) > > Just nosing around in stuff I use ... it seems like praxis' image > of the buildfarm database is not up to date. The latest build > snapshot it has is about 10 hours old: > > pgbfprod=> select max(snapshot) from build_status_log; > max > --------------------- > 2018-10-20 09:08:56 > (1 row) Good catch, update in progress. Thanks! Stephen