Log file Viewer

[00:19:40] <mongodb664> Is it a good or bad idea to store images in mongo database?

[00:21:35] <mongodb664> ah nvm, it's says it'll be loaded to RAM every time it's sent to client

[00:49:51] <meltz> hello

[00:49:57] <meltz> still no one here?

[00:51:40] <meltz> well anyway, here's the problem I'm having: http://stackoverflow.com/questions/17005313/trouble-with-automatic-failover-with-sharded-and-replicated-mongodb-setup

[10:54:15] <torbenk> Hi

[10:55:45] <torbenk> Is it OK to run a mongodb primary with v2.0 and one of the replica in 2.4?

[10:57:25] <double_p> 2.0+2.2 works, 2.2+2.4 aswell. but 2.0+2.4 -hmmm

[10:57:56] <double_p> 2.4 needs at least 2.2.3 peers, so no

[10:59:23] <torbenk> i missed some updates and need to upgrade +2.

[11:00:07] <torbenk> so i have to mix 2.0 with 2.2 to get all on 2.2 and after that combine 2.2 with 2.4 to get them all to the latest version?

[11:01:05] <ron> sounds like it.

[11:01:27] <double_p> yeah, i did that, too. 2.0->2.2.3, and then combined the 2.2.3 w/ 2.4.1

[11:02:07] <torbenk> aren't there official guides or tutorials exactly on this topic? this all seems a bit (too) tricky (for production) :-/

[11:02:28] <double_p> there's official docs on rolling upgrades

[11:04:43] <ron> torbenk: there's an easier solution. shut down the production servers, upgrade, and bring it all back up ;)

[11:12:35] <double_p> oh, and on a sidenote. dont mix repositories (like 10gen <-> epel). we had weird sideeffects there

[11:39:12] <torbenk> ron: will downtime be "only" for the time while the new version is installing or will all data be "migrated" or modified before the master can come back online again?

[11:39:31] <ron> torbenk: I was kidding...

[11:39:48] <torbenk> ron: not funny :-/

[11:39:53] <ron> A bit funny.

[13:27:24] <torbenk> using the 10gen debian packages for installing without touching the init scripts manually. Now mongodb 2.4.4 worries about NUMA. How to solve this without manually starting the mongos or modifying the init script (which will be overwritten on next upgrade) - properly?

[14:17:24] <heewa> Some ops listed by currentOp() are queries that look like they're on a collection, but the "ns" field is an empty string. Why is that?

[14:21:32] <Bartzy> What is a good naming convention for mongo hosts in a cluster ?

[14:21:34] <Bartzy> sharded cluster

[14:21:43] <Bartzy> mongos, mongod's , mongo arbiters..? How do you name them

[16:10:42] <meltz> hello

[16:11:08] <meltz> anyone here?

[16:11:32] <meltz> any ideas?: http://stackoverflow.com/questions/17005313/trouble-with-automatic-failover-with-sharded-and-replicated-mongodb-setup

[18:19:00] <HollyRain> is going to be changed the way that MongoDB handles memory?

[19:01:40] <HollyRain> is going to be changed the way that MongoDB handles memory?

[19:08:45] <heewa> I'm using mongo 2.2, which doesn't support using killOp() to kill an indexing operation. Do I have any recourse? What happens if I kill mongo during this? Will my DB be corrupt? Will it just ignore the half-made index? Do I have to repair the db (costly)?