PMXBOT Log file Viewer

Help | Karma | Search:

#mongodb logs for Sunday the 28th of April, 2013

(Back to #mongodb overview) (Back to channel listing) (Animate logs)
[13:02:53] <kelonye> Hi, model.find("...", fn) equivalent of https://gist.github.com/kelonye/5476824
[13:08:10] <Nodex> eh?
[15:33:37] <bartzy> Derick: Are you here by any chance ?
[15:34:29] <bartzy> If I remove() using multi:true and a query that has many documents - if while removing a document that matched the query was updated and now it is not matching the query, will it be removed anyway ?
[15:51:50] <Philipp|2> hi there! if I'm using SafeMode.True (C# driver) the source code states "w = 0". Does that mean SafeMode.True does not actually equal WriteConcern.Acknowledged?
[16:08:48] <Derick> Philipp|2: it should - where do you see the w=0?
[16:36:25] <apocn> How can I return this document like an assoc. array? http://ideone.com/g1hOOW
[16:38:00] <Philipp|2> Derick: well it is the default when using the constructor of SafeMode without a w parameter and that is used by the static True instance. since it gets frozen, too, it should not be able to change anywhere else?
[16:40:16] <Derick> I don't know the driver well enough, but you are using MongoClient, not?
[17:26:43] <bartzy> hey
[17:26:46] <bartzy> If I remove() using multi:true and a query that has many documents - if while removing a document that matched the query was updated and now it is not matching the query, will it be removed anyway ?
[17:27:51] <ron> all hail database locks!
[17:28:02] <bartzy> ron: Do you have an idea? :)
[17:28:19] <bartzy> the docs say that the remove will be interleaving with other writes
[17:28:27] <bartzy> but does that mean that what I described can happen ?
[17:28:43] <ron> not sure, honestly.
[18:15:19] <Derick> bartzy: you asked a question earlier?
[18:16:44] <bartzy> Derick: Yeah
[18:16:46] <bartzy> If I remove() using multi:true and a query that has many documents - if while removing a document that matched the query was updated and now it is not matching the query, will it be removed anyway ?
[18:16:52] <bartzy> sorry , no need for multi:true
[18:17:28] <bartzy> but if the remove removes 10,000 documents, which takes 10 seconds to complete. The 9,999th document gets updated after 1 second, which gets it out of the query to remove. what will happen ?
[18:18:01] <Derick> bartzy: it won't be removed then
[18:18:21] <bartzy> Derick: So each document is being evaluated by itself when it is removed ?
[18:18:22] <Derick> but it's difficult to predict as it depends on how the result set is traversed
[18:18:25] <Derick> yes
[18:18:39] <bartzy> Doesn't it get a cursor, and just remove() every document in that cursor ?
[18:19:01] <bartzy> Derick: Are you sure ? I don't know how to test this since it happens too fast. :\
[18:19:12] <bartzy> What is the point of $isolated then, in the docs...
[18:21:11] <Derick> the cursor is not fixed
[18:21:21] <Derick> it's just a pointer going through a btree
[18:22:38] <Derick> you didn't mention $isolated before ;-)
[18:23:07] <Derick> sounds like $isolated would take out a full write lock for the whole duration... which is a bad idea
[18:58:13] <bartzy> Derick: What's the point in $isolated if the behavior of the cursor is like you said ?
[19:12:59] <bartzy> Derick: ?
[19:23:24] <Derick> bartzy: sorry, it was dinner time
[19:23:37] <Derick> bartzy: $isolated changes the behaviour from what I mentioend
[19:39:26] <bartzy> Derick: But why lock the DB if it does not change the behavior ?
[19:39:46] <bartzy> I mean, why remove with $isolated:true, if mongo evaluates each document in the cursor again, before doing the remove ?
[21:35:15] <gazarsgo> is there a quick & dirty way to see what size the largest object is per collection ?
[22:43:02] <tyler_> whats the best way to replace drive if im running out space on my mongo RAID10
[22:43:18] <tyler_> out of*
[22:44:41] <tyler_> anyone know if its ok to add data to a new raid10 drive beforeit syncs all the way?
[22:49:05] <gazarsgo> are you running 1 node or 3 or ?
[22:55:23] <tyler_> gazarsgo: 1 node right now
[22:56:51] <tyler_> from what i gather it would involve setting up another node and making that the master
[22:56:55] <tyler_> im a mongo newb
[23:16:01] <gazarsgo> i'm a mongo noob too but i think you need 3 nodes to make another one master
[23:27:41] <skot> tyler_: are you using lvm or softraid?
[23:29:00] <tyler_> skot: not sure. i made the disk with mdadm, formatted it and then mounted it