Log file Viewer
#mongodb logs for Tuesday the 1st of September, 2015
[06:41:09] <aps> I locked my RS secondary using fsyncLock and now I'm unable to Unlock it. db.fsyncUnlock is just stuck. What to do? :/
[06:43:41] <stekla> hi guys. i tried to restore my mongo db with mongorestore, but i got an error: Failed: error connecting to db server: no reachable servers. How i can fix this?
[06:43:57] <aps> now, doing mongo --eval "db.fsyncUnlock()" does nothing, even tried db.fsyncUnlock() by going in mongo shell
[06:52:23] <aps> joannac: this is what I get, then it's stuck https://www.irccloud.com/pastebin/JjQJdHAD/
[06:53:18] <aps> joannac: it's not the same shell actually, I had to re-login. Could that be a problem?
[06:54:46] <joannac> aps: fsyncLock blocks all writes. writes takes precendent over reads. one write when fsyncLocked means no reads will occur. that includes auth
[06:57:36] <joannac> stekla: please follow the syntax in the docs http://docs.mongodb.org/manual/reference/program/mongorestore/
[06:58:25] <joannac> assuming you want to restore to a db called "my_dump" and the actual dumpfiles are in the current directory in a folder called "my_dump"
[07:02:54] <stekla> but i used mongorestore my_dump -d my_dump command on another computer and it worked pretty well.
[07:31:52] <tagrudev> hello guys I am having this issue with mongo and a request for removing a string from a nested array
[07:33:37] <Boomtime> tagrudev: your use of $ positional operator has no match in the query portion, you matched on an exact match of _id - there is no array position for $ to represent
[07:36:27] <Boomtime> tagrudev: if you expect to match _id inside the modifications array then your query should be {"modifications._id":id}
[07:38:41] <Boomtime> you need to include the user match in the query portion so the $ has something to reference
[07:40:48] <Boomtime> to be honest, i think you'd be better off doing the modification on the client - the result is potentially complex - either that or you would need to run the update multiple times to ensure it hit every match
[07:41:26] <Boomtime> an update with a predicate only matches once - it doesn't keep matching and applying further updates to the same document - it makes exactly one change
[08:28:35] <alaee> Hi guys, is a result of .explain('executionStats) "executionTimeMillisEstimate" : 0 something good?
[09:52:10] <aps> Suppose all (two) mongo members in a RS are down for sometime, then they come back but one is in "RECOVERING" state and other is "SECONDARY". How do I make one of them primary?
[09:54:15] <aps> Derick: it seems to be stuck in RECOVERING state for quite some time now. How do I make sure there is progress in recovery process?
[09:54:59] <aps> okay, there's no primary left right now. Does a member still recover when there is no primary?
[09:56:47] <aps> I see lots of these - [rsHealthPoll] replset info x.x.x.x:2701 heartbeat failed, retrying - in the log @Derick
[09:58:20] <Derick> which IP is being shown? is that the IP of the other node? and is the port really 2701?
[10:00:02] <aps> Derick: yes, both IP and port are correct. I think the issue is that I was trying to add auth and I have screwed it somehow
[10:14:55] <aps> Derick: yeah, that was the issue. Related question, is it possible to enable auth for just one collection and not on db level?
[11:58:16] <yopp> And working set can't fit in the memory, but we're not using all the data on most popular app endpoint, only relatively recent
[12:00:15] <yopp> As far I understand, in this case mongo still be forced to go to the disk to crawl over the full index, right?
[12:02:01] <yopp> I mean, in the best case scenario, did index lookup require all the index, or just a part?
[12:05:27] <aps> how do I do opposite of rs.initiate? I.E. I want to remove replica set config from mongo member and make it standalone
[12:38:39] <arussel> what is the recommanded way to install mongo on ubuntu 15.04 ? using the ubuntu package or mongo 14.04 package ?
[12:51:30] <deathanchor> always use the packages from the mongodb site, I don't trust the repos to keep up to date.
[12:51:50] <StephenLynx> yeah, but the packages from mongod don't support anything over 14.04 I guess
[14:39:55] <KekSi> any word on the java driver splitting in 2.x and 3.x? is 2.x going to be developed in the future or does it run out soon?
[14:50:11] <fartface> I'm trying to model an ice-hockey team into a MongoDB database, but I'm a bit confused as to how to access nested properties--I've got a pastebin up at http://pastebin.com/3bBL9abf, does anyone know what the best way to handle this would be?
[14:53:50] <StephenLynx> the problem with that insert is that you are not setting the seasons array for the player before inserting
[14:55:31] <fartface> Would that make more sense? I had originally thought about doing it that way, but then in reading around, it sounded like it was more advisable to have it embedded in the related part (the player), but if it's totally fine to have a "Seasons" collection, then I can just do that and insert the player ID
[14:56:03] <fartface> So then I could call it like Seasons.findOne({name: '2015 - 2016', player: (some ID)})
[14:56:36] <StephenLynx> in your case, I don't see much of a difference, since there won't be many seasons per player
[14:56:48] <Pinkamena_D> If you don't embed you will have to usually query twice, which looks worse in the code but I have found that there is not much of a performance impact.
[14:57:45] <StephenLynx> having it on a separate collection will impose the following limitation on you:
[14:57:54] <fartface> I wasn't sure whether it'd make more sense to do a second query or to do a loop to get the right season. A second query felt "cleaner", but wasn't sure of performance impact
[14:58:03] <StephenLynx> when you are listing players, it won't work so well if you wish to list them along with season data.
[14:58:40] <fartface> Ah, and that I do. I want to have a "roster" that'll output each players information with their latest season stats
[14:59:09] <StephenLynx> I suggest you embed this latest season stats and create a separate collection for past season data.
[15:01:02] <fartface> Actually that's a pretty good idea. I'll have to think a bit more about how to handle "seasons", in that, let's say that a player plays this season, but not the next. When I create a new season, I guess I'll have some application logic to go through each of the players and push the current season data to an 'archive_season' collection and then null out the 'current' season data
[15:01:53] <StephenLynx> I have a collection with hourly stats and I replicate the hourly stat at given times to the board document.
[15:02:07] <fartface> So that when I go to pull up the current roster, it wouldn't pull up players that played 3 seasons ago and never again or something like that.
[15:14:12] <brianV> Hi all. I'm trying to project hardware requirements for a Mongo deployment. We have about 120gb of data, and we are planning on going to Amazon EC2 and use the Mongo Cloud Manager
[15:14:33] <brianV> what type of instances are normally recommended for Mongo? Better to look at multiple small instances, or one large instance?
[15:17:36] <brianV> StephenLynx: what would you recommend? The rest of our infrastructure is on EC2 already, so we benefit from good internal latency
[15:23:51] <cq_dev> Hello, is it known whether or not Mongo is going to add "$hint" to the aggregation framework?
[15:24:32] <cq_dev> I have an issue whether I'm forced to using "$geoNear" which forces me to use one geospatial index for the collection.
[15:33:48] <yopp> but generally you need to be sure that your indexes and most used data can fit in memory
[15:34:47] <yopp> My advice: try to push the data with your real indexes in the mongodb and then try to query it, to decide on indexes
[15:37:12] <brianV> we have like 500mb of 'static' data, then everything else is records keyed by users, very shardable by userID
[17:05:18] <leev> I set the wrong shard key on a collection, so tried to stop the balancer and drop the collection, but it has a lock from one of the mongos instances.
[17:41:59] <cheeser> that'd be my guess. but given how documented that is, calling it "bad" seems ... bad.
[19:15:31] <diegoaguilar> hello, can someone help me with this pagination issue? http://stackoverflow.com/questions/32339144
[19:20:37] <diegoaguilar> Ok, I thought so after this: http://blog.mongodirector.com/fast-paging-with-mongodb/
[19:25:21] <diegoaguilar> dethanchor I thought on db.lineuppointsrecord.find({round: 0, ranking: {$gt: 6}}).sort({ranking: 1}).limit(3)
[19:25:55] <diegoaguilar> can I get performance optimization with a composed {round: 1, ranking: 1} index?
[20:17:00] <deathanchor> I'm getting an assert error with this: var dbo = connect("mongodb://localhost:27017/users?readPreference=secondary");
[20:39:00] <BaNzounet> Hey guys, I've a collection of 1M records, I want to rename a field let's say fooId : { $oid: "aabbcc" } to barId: { $oid : "aabbcc" } what do you advice? Can I use $rename?
[23:27:40] <MacWinne_> joannac, just periodically checking with you if wired tiger is mature enough for your production use..