# Inserted\_at seems buggy in production mode? **URL:** \ **Category:** Rails\ **Created:** [March 16, 2006, 12:37pm UTC](https://www.ruby-forum.com/t/inserted-at-seems-buggy-in-production-mode/82507 "2006-03-16T12:37:48Z")\ **Posts on this page:** 4\ **Page:** 1
**Author:** ![jeroen](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/j/bbe5ce/32.png) [@jeroen](https://www.ruby-forum.com/u/jeroen)\ **Post date:** [March 16, 2006, 12:37pm UTC](https://www.ruby-forum.com/t/inserted-at-seems-buggy-in-production-mode/82507/1 "2006-03-16T12:37:48Z")
Hi, I’ve got a table with an inserted\_at column. When I insert a record it automatically injects the current date/time, all is well. Then when I insert a second record, it uses the exact same timestamp that is used in the first record. This behavior only seems to occur in production mode. When I restart the server, the problem goes away but then immidetatly occurs again on the second insert. I’m using postgresql 8.1 and I have my default value for the inserted\_at column set to now(). Any ideas?? Jeroen — below is some log output ---- Processing MatchController#comment (for 192.87.30.169 at 2006-03-16 12:23:41) [POST] Parameters: {“mc”=\>{“body”=\>“testtest”}, “commit”=\>“Een moment …”, “action”=\>“comment”, “id”=\>“159”, “controller”=\>“match”} Match Load (0.001001) SELECT \* FROM matches WHERE (matches.id = ‘159’) LIMIT 1 SQL (0.000319) BEGIN PK and serial sequence (0.017476) SELECT attr.attname, name.nspname, seq.relname FROM pg\_class seq, pg\_attribute attr, pg\_depend dep, pg\_namespace name, pg\_constraint cons WHERE seq.oid = dep.objid AND seq.relnamespace = name.oid AND seq.relkind = ‘S’ AND attr.attrelid = dep.refobjid AND attr.attnum = dep.refobjsubid AND attr.attrelid = cons.conrelid AND attr.attnum = cons.conkey[1] AND cons.contype = ‘p’ AND dep.refobjid = ‘match\_comments’::regclass MatchComment Create (0.002280) INSERT INTO match\_comments (“body”, “match\_id”, “player\_id”, “inserted\_at”) VALUES(‘testtest’, 159, 14, ‘2006-03-16 12:22:55’) SQL (0.000684) SELECT currval(‘public.match\_comments\_id\_seq’) SQL (0.034973) COMMIT Redirected to [http://tcuithof.laddertennis.nl/match/show/159](http://tcuithof.laddertennis.nl/match/show/159) Processing MatchController#comment (for 192.87.30.169 at 2006-03-16 12:24:57) [POST] Parameters: {“mc”=\>{“body”=\>“test”}, “commit”=\>“Een moment …”, “action”=\>“comment”, “id”=\>“159”, “controller”=\>“match”} Match Load (0.001003) SELECT \* FROM matches WHERE (matches.id = ‘159’) LIMIT 1 SQL (0.000327) BEGIN MatchComment Create (0.000929) INSERT INTO match\_comments (“body”, “match\_id”, “player\_id”, “inserted\_at”) VALUES(‘test’, 159, 14, ‘2006-03-16 12:22:55’) SQL (0.000448) SELECT currval(‘public.match\_comments\_id\_seq’) SQL (0.036623) COMMIT Processing MatchController#comment (for 192.87.30.169 at 2006-03-16 12:26:00) [POST] Parameters: {“mc”=\>{“body”=\>“dsasda”}, “commit”=\>“Een moment …”, “action”=\>“comment”, “id”=\>“159”, “controller”=\>“match”} Match Load (0.001013) SELECT \* FROM matches WHERE (matches.id = ‘159’) LIMIT 1 SQL (0.000307) BEGIN MatchComment Create (0.000991) INSERT INTO match\_comments (“body”, “match\_id”, “player\_id”, “inserted\_at”) VALUES(‘dsasda’, 159, 14, ‘2006-03-16 12:22:55’) SQL (0.000462) SELECT currval(‘public.match\_comments\_id\_seq’) SQL (0.034870) COMMIT ---
**Author:** ![jeroen](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/j/bbe5ce/32.png) [@jeroen](https://www.ruby-forum.com/u/jeroen)\ **Post date:** [March 16, 2006, 12:46pm UTC](https://www.ruby-forum.com/t/inserted-at-seems-buggy-in-production-mode/82507/2 "2006-03-16T12:46:58Z")
Jeroen H. wrote: > Hi, > > I’ve got a table with an inserted\_at column. When I insert a record it > automatically injects the current date/time, all is well. > Then when I insert a second record, it uses the exact same timestamp > that is used in the first record. Okay I should have searched for this before I posted [http://dev.rubyonrails.org/ticket/2257](http://dev.rubyonrails.org/ticket/2257) More logic to move out of the database. I may as well start using SQLite! 😉 Jeroen ---
**Author:** ![jeroen](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/j/bbe5ce/32.png) [@jeroen](https://www.ruby-forum.com/u/jeroen)\ **Post date:** [March 16, 2006, 12:47pm UTC](https://www.ruby-forum.com/t/inserted-at-seems-buggy-in-production-mode/82507/3 "2006-03-16T12:47:06Z")
I think your problem is related to this ticket: [http://dev.rubyonrails.org/ticket/2257](http://dev.rubyonrails.org/ticket/2257) In production mode default value is cached right after the first query for table columns metadata, so it will be the same on subsequent requests. You could rename ‘inserted\_at’ to ‘created\_at’ and use Rails magic of ActiveRecord::Timestamp mixin (active\_record/timestamp.rb) ---
**Author:** ![jeroen](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/j/bbe5ce/32.png) [@jeroen](https://www.ruby-forum.com/u/jeroen)\ **Post date:** [March 16, 2006, 12:59pm UTC](https://www.ruby-forum.com/t/inserted-at-seems-buggy-in-production-mode/82507/4 "2006-03-16T12:59:39Z")
Lugovoi N. wrote: > I think your problem is related to this ticket: > [http://dev.rubyonrails.org/ticket/2257](http://dev.rubyonrails.org/ticket/2257) > In production mode default value is cached right after the first query > for table columns metadata, so it will be the same on subsequent > requests. > > You could rename ‘inserted\_at’ to ‘created\_at’ and use Rails magic of > ActiveRecord::Timestamp mixin (active\_record/timestamp.rb) Thanks! Aha I thought it should be called inserted\_at, but created\_at is the special column name. My bad! So if I do that, what should I use as my column definition? NOT NULL? And no default value? Jeroen