Обсуждение: bgwriter_lru_maxpages limits in PG 10 sample conf
Hello We increased bgwriter_lru_maxpages limit in 10 release [1]. Docs now are changed correctly but in REL_10_STABLE postgresql.conf.samplewe still have comment "0-1000 max buffers written/round". Master (and REL_11_STABLE) was updated later in 611fe7d4793ba6516e839dc50b5319b990283f4f, but not REL_10. I think we needbackpatch this line. * http://postgr.es/m/f6e58a22-030b-eb8a-5457-f62fb08d701c@BlueTreble.com regards, Sergei
Вложения
On Wed, Feb 20, 2019 at 5:53 AM Sergei Kornilov <sk@zsrv.org> wrote: > We increased bgwriter_lru_maxpages limit in 10 release [1]. Docs now are changed correctly but in REL_10_STABLE postgresql.conf.samplewe still have comment "0-1000 max buffers written/round". > Master (and REL_11_STABLE) was updated later in 611fe7d4793ba6516e839dc50b5319b990283f4f, but not REL_10. I think we needbackpatch this line. I'm a bit reluctant to whack postgresql.conf around in back-branches because sometimes that makes funny things happen when somebody upgrades, e.g. via RPM. I don't remember exactly what happens but I think typically either the new file overwrites the existing file which gets moved to something like postgresql.conf.rpmsave or the new file is written into postgresql.conf.rpmnew instead of the original location. I don't think it's worth making stuff like that happen for the sake of a comment. -- Robert Haas EnterpriseDB: http://www.enterprisedb.com The Enterprise PostgreSQL Company
Hello > I'm a bit reluctant to whack postgresql.conf around in back-branches > because sometimes that makes funny things happen when somebody > upgrades, e.g. via RPM. If i remember correctly both deb and rpm packages will ask user about config difference. But good point, comment change is too small difference. I am a bit late, good time for such change was before last minorrelease (we add data_sync_retry and config was changed anyway). regards, Sergei
On Wed, Feb 20, 2019 at 11:54:25PM +0300, Sergei Kornilov wrote: > Hello > > > I'm a bit reluctant to whack postgresql.conf around in back-branches > > because sometimes that makes funny things happen when somebody > > upgrades, e.g. via RPM. > > If i remember correctly both deb and rpm packages will ask user about config difference. > But good point, comment change is too small difference. I am a bit late, good time for such change was before last minorrelease (we add data_sync_retry and config was changed anyway). The other issue is that you will change share/postgresql.conf.sample, but not $PGDATA/postgresql.conf until initdb is run, meaning if someone diffs the two files, they will see differences that they did not make. Making defaults more accurate is not worth that confusion. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com + As you are, so once was I. As I am, so you will be. + + Ancient Roman grave inscription +
Hello postgresql.conf.sample was changed recently in REL_10_STABLE (commit ab1d9f066aee4f9b81abde6136771debe0191ae8) So config will be changed in next minor release anyway. We have another reason to not fix bgwriter_lru_maxpages comment? regards, Sergei
On Thu, Feb 28, 2019 at 10:28:44AM +0300, Sergei Kornilov wrote: > Hello > > postgresql.conf.sample was changed recently in REL_10_STABLE (commit ab1d9f066aee4f9b81abde6136771debe0191ae8) > So config will be changed in next minor release anyway. We have another reason to not fix bgwriter_lru_maxpages comment? Frankly, I was surprised postgresql.conf.sample was changed in a back branch since it will cause people who diff $PGDATA/postgresql.conf with share/postgresql.conf.sample to see differences they didn't make. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com + As you are, so once was I. As I am, so you will be. + + Ancient Roman grave inscription +
On 2019-Mar-04, Bruce Momjian wrote: > On Thu, Feb 28, 2019 at 10:28:44AM +0300, Sergei Kornilov wrote: > > Hello > > > > postgresql.conf.sample was changed recently in REL_10_STABLE (commit ab1d9f066aee4f9b81abde6136771debe0191ae8) > > So config will be changed in next minor release anyway. We have another reason to not fix bgwriter_lru_maxpages comment? > > Frankly, I was surprised postgresql.conf.sample was changed in a back > branch since it will cause people who diff $PGDATA/postgresql.conf with > share/postgresql.conf.sample to see differences they didn't make. I think the set of people that execute diffs of their production conf file against the sample file to be pretty small -- maybe even empty. If you're really interested in knowing the changes you've done, you're more likely to use a version control system on the file anyway. -- Álvaro Herrera https://www.2ndQuadrant.com/ PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
On Tue, Mar 5, 2019 at 12:24:14AM -0300, Alvaro Herrera wrote: > On 2019-Mar-04, Bruce Momjian wrote: > > > On Thu, Feb 28, 2019 at 10:28:44AM +0300, Sergei Kornilov wrote: > > > Hello > > > > > > postgresql.conf.sample was changed recently in REL_10_STABLE (commit ab1d9f066aee4f9b81abde6136771debe0191ae8) > > > So config will be changed in next minor release anyway. We have another reason to not fix bgwriter_lru_maxpages comment? > > > > Frankly, I was surprised postgresql.conf.sample was changed in a back > > branch since it will cause people who diff $PGDATA/postgresql.conf with > > share/postgresql.conf.sample to see differences they didn't make. > > I think the set of people that execute diffs of their production conf > file against the sample file to be pretty small -- maybe even empty. If > you're really interested in knowing the changes you've done, you're more > likely to use a version control system on the file anyway. Well, if this is true, then we should all agree to backpatch to postgresql.conf.sample more often. -- Bruce Momjian <bruce@momjian.us> http://momjian.us EnterpriseDB http://enterprisedb.com + As you are, so once was I. As I am, so you will be. + + Ancient Roman grave inscription +
Hello Well, actually we change postgresql.conf.sample in back-branches. Recently was yet another commit in REL_10_STABLE: fea2cab70de8d190762996c7c447143fb47bcfa3 I think we need fix incosistent comment for bgwriter_lru_maxpages regards, Sergei