Log file Viewer
#mongodb logs for Thursday the 4th of September, 2014
[02:06:05] <ohshitdacops> So I have a strange problem. I would provide logs, but there are none to provide
[02:06:53] <ohshitdacops> So I have MongoDB installed on one of my EL7 development environments and have noticed that whenever I restart the box, mongodb refuses to start and requires a full reinstall
[02:08:10] <ohshitdacops> The little info I can provide: [BOX: "CentOS 7", PACKAGE: "yum, via mongodb repo", ERROR: "failed state, no logs"]
[05:26:53] <royaa1> Which style of mongoose query should I be using? The first or the second? https://gist.github.com/anonymous/922459bdcac5482d3132
[05:32:33] <catNode> Boomtime: I prefer the second one, but for some reason I have in my head that's an old way of doing it? I must've read that in documentation or on SO somewhere?
[05:34:07] <Boomtime> the second tends to be eaier to read, construct, and maintain - the first one is more like what the server actually gets sent
[05:35:29] <Boomtime> not really, the work to convert the second one is done on the client - but if that tiny amount of work affects you then you shouldn't be running a javascript interp in the first place
[09:13:15] <krion> for those who follow my trouble, even with cleaning all /var/lib/mongo dbs, i still have trouble
[09:13:41] <krion> when i try to shard a database, it say "{ "ok" : 0, "errmsg" : "already enabled" }"
[09:16:31] <rspijker> Well, sharding being enabled is stored in the cfg servers iirc. The mongos processes do cache the config, but you can invalidate that
[09:26:02] <rspijker> can you pastebin the output of that command and the command you are using to enable sharding?
[09:30:29] <rspijker> well.. it says { "_id" : "mydatabase", "partitioned" : true, "primary" : "rs1" } in sh.status()….
[10:39:42] <A13> hi there, I got a simmple question: can I copy the data files (from /data/db) to another server and start mongod there? will it work? I cannot start a damaged installation of mongodb to do a proper export..
[11:00:00] <tehmoon> does anybody know why the {files_id: "hashed"} index for gridfs does not work on a sharded cluster ? v2.6.2
[12:27:20] <shambat> I have data stored with entries like this: http://bpaste.net/show/2f6fd25f8306 -> How can I make a query to find entries where "sources" array is larger than 1. I have tried this, but it returns nothing: db.sigs.find( {sources : {$exists:true}, $where:'this.sources.length>1'} )
[13:03:28] <rspijker> if you only want larger than 1, you can also use {“sources.1” : {$exists:true}}
[13:04:28] <rspijker> caveat is when you could have sources that aren’t arrays but documents and could contain a field named 1…
[13:26:56] <remonvv> Hm, detail perhaps, but does anyone know why the shell prints NumberLong(1084752864) for numbers < 2^32 and NumberLong("10665270000") so with quotes for numbers > 2^32
[13:30:19] <jordana> remonvv: I think JS only has a type that supports 8byte ints so it's a workaround
[13:32:38] <remonvv> The single reason all languages with a best guess type should have been retired decades ago ;)
[15:25:15] <cheeser> tscanausa: http://www.mongodb.com/press/mongodb-delivers-multi-petabyte-data-store-option-adobe-experience-manager
[15:25:57] <lha13> Hello, I've got the following stackoverflow question (http://stackoverflow.com/questions/25666187/mongodb-nested-group/25667256) and I 'm looking at
[15:26:28] <lha13> I'm looking at limiting the amount of results to 5 per array. How would I go about doing this? Any advice would be great thans!
[16:03:22] <cheeser> when you're talking about data in the petabyte range, how specific do you *really* need to be?
[16:18:53] <oli47> well basically I am trying to do an update of some documents, and I need to match only the ones that do NOT contain a specific value in an array value
[17:21:53] <doug1> Why would I get, on the router, "can't authenticate to <some-cfg-server> as internal user, error: { ok: 0.0, errmsg: "auth failed", code: 18 }" ?
[18:43:44] <doug1> Why would I get, on the router, "can't authenticate to <some-cfg-server> as internal user, error: { ok: 0.0, errmsg: "auth failed", code: 18 }" ?
[18:50:15] <harttho> Where is the configuration for the mongos connection ports that would result in taking up 9099?
[18:59:16] <harttho> Note: this is on a server with mongos running connecting to an external mongo server
[19:11:21] <doug1> Why would I get, on the router, "can't authenticate to <some-cfg-server> as internal user, error: { ok: 0.0, errmsg: "auth failed", code: 18 }" ?
[20:12:42] <skot> doug1, check to make sure you are using the same keyfile or x509 cert for both the mongos and config server.
[20:14:01] <skot> harttho: that is how ephemeral ports work when a process has to make outgoing connections; it is expected.
[20:18:26] <mattkremer> Alright, I'm in the middle of a meeting with my team and there are some crazy ideas floating around regarding speed of lookup with indexes. Would hashing the values of data in MongoDB actually increase the speed of lookups?
[20:20:10] <mattkremer> Yes, let's say that an index already exists. Are queries against that index going to be different if the data was stored as full string or as hash when it was initially inserted and indexed.
[20:20:51] <cheeser> you'd have to query using that hashed value and at that point, you're just doing an index lookup either way
[20:20:57] <mattkremer> That's what I thought...since the data is already indexed in an actual lookup, that the query actually won't differ in speed.
[20:21:29] <cheeser> now, if you're talking strings, larger strings are only indexed up to a certain length, iirc.
[20:23:15] <shadfc> I'm getting an OperationFailure "can't get runner for query". Trying to upload a file to gridfs using pymong. Mongdb 2.4 currently. Any ideas?
[20:23:24] <mattkremer> Alright, so if I know our data isn't going to be over 1KB, then it's not really going to matter at all?
[20:48:34] <skot> mattkremer, there is going to some measurable improvement with smaller strings since they will result in a smaller indexes and faster lookups.
[21:02:39] <skot> It is a mongod option. It would prevent you from having queries unless they use an index, and that error message.
[21:04:23] <shadfc> I'm seeing it in two mongo installations. the common options between them are --noprealloc --smallfiles
[21:06:56] <skot> shadfc: Can you run this from the shell/app getting the error? db.adminCommand("getCmdLineOpts")
[21:09:21] <skot> Can you post a paste with the query, output, list of indexes and stats for that collection?
[21:14:43] <shadfc> OperationFailure: command SON([('filemd5', ObjectId('5408d55cbaba325b82964932')), ('root', u'fs')]) failed: exception: Can't get runner for query { files_id: ObjectId('5408d55cbaba325b82964932'), n: { $gte: 0 } }
[21:27:12] <akp> is there anyway to just return it as a list or something so i can get a count from it and other things?
[21:30:50] <skot> I'm no php programmer, so I can't help there but would assume this would be a simple thing. In python it is just list(cursor), but php, don't know.
[21:35:24] <tscanausa> akp: you have to iterate the entire cursor. if you want a count call the count function with the query
[21:37:21] <tab1293> is there a faster way to convert a field type other than looping through every document with a foreach?
[21:38:39] <tab1293> I have about 350,000 docs in the collection. the foreach has been running for like 10 minutes
[21:39:42] <skot> tab1293: Do you mean the same type, not same value. String is the type, "13", "15" or different values of strings.
[21:42:17] <skot> tab1293: yes, but you can update just that field, and run multiple threads to do the work since your client is probably the slow part.
[21:42:34] <skot> (also, just retrieving the field you want to change will reduce the network traffic)
[21:46:34] <skot> yes, and don't use save, use db.update({_id:x._id}, {$set:{tpb_id:new NumberInt(x.tpb_id)}})
[21:48:04] <tab1293> am I safe to open a new shell and run the modified command while this once is running?
[21:50:00] <skot> You can also use the bulk api to get a bunch, group them, and send them back as a list of individual updates.
[22:13:40] <harttho> skot: is there a way to restrict which ports are being used for the outgoing connections?
[22:15:20] <harttho> Using Ubuntu 12.04 if you happen to know off the top of your head. We've run into some port collision of processes we have on the same ports and get eaddrinuse
[22:18:22] <skot> I searched the internet to return you these results: cat /proc/sys/net/ipv4/ip_local_port_range , echo "X Y" > /proc/sys/net/ipv4/ip_local_port_range
[22:33:13] <freeone3000> I'm trying to shard *some* databases on the server, but not others. I've successfully (probably) sharded a database called "company-prod", but I'm attempting to *not* shard "company-config". However, mongo gives me https://gist.github.com/freeone3000/9cd6b4089246707dd91e . What am I doing wrong?
[22:59:56] <tab1293> joannac: i am connecting to the server over ssh and want the operation to run when I logout
[23:36:16] <unholycrab> i want to add a new empty SECONDARY to a replica set, let it sync up all the data and join the replica set.... but i want to control when the election happens
[23:36:33] <unholycrab> ie: if i just let it go, it will force an election as soon as its done copying, whenever that happens to be
[23:42:53] <joannac> unholycrab: no, it won't force an election at all, assuming it doesn't ahvea higher priority than the current primary
[23:45:19] <joannac> well, any version 2.4+. I cdon't have enough experience with 2.2 and below to say for sure
[23:50:32] <freeone3000> I'm trying to shard *some* databases on the server, but not others. I've successfully (probably) sharded a database called "company-prod", but I'm attempting to *not* shard "company-config". However, mongo gives me https://gist.github.com/freeone3000/9cd6b4089246707dd91e . What am I doing wrong?