Log file Viewer
#mongodb logs for Sunday the 3rd of July, 2016
[03:18:25] <GothAlice> http://api.mongodb.com/python/current/api/bson/binary.html#bson.binary.UUID_SUBTYPE < this is silly
[03:20:52] <GothAlice> http://api.mongodb.com/python/current/api/bson/codec_options.html#bson.codec_options.CodecOptions < default here
[03:31:22] <Boomtime> @GothAlice: i don't see any jira ticket even talking about it, you could always raise one to request the default be changed -> https://jira.mongodb.org/browse/PYTHON
[04:10:54] <kamoricks> Having a replication problem. https://gist.github.com/freeone3000/ff2a169849c951f7472a7f6a2ffacb28 is my current rs.status(). As you can see, one of the nodes hs been in "initializing" for nearly three minutes. When I check rs.status() from that node, it claims "replSet error loading set config (BADCONFIG)"
[04:15:15] <Boomtime> @kamoricks: gist the replset config too, from the primary (rs.conf()) and from the stuck member if it will do it, otherwise check the member log file as it should print something about why it doesn't like the config
[04:16:26] <kamoricks> Boomtime: https://gist.github.com/freeone3000/b2aab19d3c3c66b1c0e32527063d860e
[04:16:44] <kamoricks> Boomtime: It appears to think it's in the 'bbia' replset despite mongod.conf saying it's in the bbia-2 replset. This *could* cause issues.
[04:19:00] <Boomtime> where did it get that config from? it certainly didn't get it from any other member
[04:19:37] <Boomtime> shut it down and delete the dbpath, if you want a new member in a replset the dbpath must be empty or it will halt (to protect any possible data that might be present)
[04:20:57] <kamoricks> Boomtime: Awesome, it's now in STARTUP2, doing full recovery (but in the dbset). Thanks.
[04:24:09] <Boomtime> that advice is only true if you have a high availability replica-set and you know there is nothing you want in the dbpath of that particular member - then it just happens to be the simplest course of action, specifically; let the other members heal the dysfunctional one
[15:09:21] <OSInet> Hello. I can't seem to make use of the MongoDB stream context with the new (phongo) driver, although it used to work in the legacy driver, and AIUI PHPC-30 seems to imply it should work (OTOH PHPLIB-38 seems to imply it doesn't)
[15:16:51] <OSInet> I keep seeing references to phongo.debug_log as a php.ini setting in the extension source, but it doesn't seem to produce any output
[17:19:42] <GothAlice> Ouchies; that's new enough to have some interesting outstanding bugs. PHP 7 memory leaks, inability to iterate tailable cursors, redundant SSL/TLS negotiation, segfault during shutdown, broken SASL on Mac, recursive bsonSerialize segfaults, and a not insignificant amount of bad/confusing error reporting. :(
[17:36:25] <OSInet> GothAlice: I seem to have less problems when using the PECL 1.1.7 version than the 1.1.7 from the Ondrej PPA, which corrupts the heap (on Ubuntu 12.04 LTS).
[17:38:29] <OSInet> Looks like it should be possible to hijack the output from mongodb.debug and filter it to obtain something useful : only keep lines matching stream: TRACE and not matching ENTRY|EXIT and you essentially get the readv/writev calls and the hexdumped BSON messages.
[18:04:42] <bros> Do I need to change my writeConcern and readPreference settings to anything special for a production environment to avoid the common pitfalls of MongoDB?
[18:11:39] <jamiel> Hi all, just trawling through the docs trying to find how I would I create an index to cover queries by array index position - ie. db.insert({"foo": [1,2] }); db.find({"foo.0": 1}); but to no avail. Is this possible? Thanks.
[20:51:43] <Anon4254> So, I want to install MongoDB for Windows 7, and I'm concerned by this warning in the documentation: "IMPORTANT: If you are running any edition of Windows Server 2008 R2 or Windows 7, please install a hotfix to resolve an issue with memory mapped files on Windows." That hotfix is no longer available (MS gives a 404). Is the hotfix still necessary for current versions of MongoDB?
[20:56:20] <jamiel> Looks like I can create the index with the position { "foo.0": 1 } to cover that query.
[20:58:11] <GothAlice> Anon4254: I do not know if that hotfix made it to general deployment on Windows Update for Windows 7, so for now I would assume "yes, it's still needed, darn Microsoft and their bit rot". It's a fix for a flaw in Windows, not a fix for MongoDB it self.
[21:00:32] <Anon4254> @GothAlice: Thanks for the info. So, in practice, Windows 7 can't run MongoDB? Are there any workarounds or anything?
[21:02:11] <GothAlice> https://support.microsoft.com/en-us/hotfix/kbhotfix?kbnum=2731284&kbln=en-us < I won't agree to anything to get past this point to check if the problem is there, but, it 404's after this?
[21:02:40] <Anon4254> that page redirects to "https://support.microsoft.com/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/en/hotfix/kbhotfix?kbnum=2731284&kbln=en-us". MS's site is messed up.
[21:04:55] <GothAlice> Anon4254: Check the "file information notes" section of the KB article, Google around (or use site search) for the appropriate manifest file for your version.
[21:06:20] <GothAlice> Anon4254: Or your browser is sending a bad Accept-Language header. I will agree on the site bit, I find it rarely actually works for me. Googling the manifests and a good AV scanner are your best bet, sadly.
[21:07:31] <GothAlice> javawolfpack: If the node is standalone, aye, just copy over the config and data directory, spin it up. If it's a member of a replica, demote the replica, transfer the data, and add back to the set under its new name.
[21:08:43] <javawolfpack> GothAlice: Thanks, it’s currently up on red hat’s openshift, so not sure how much access I have to the files :-/
[21:08:51] <Anon4254> @GothAlice: Huh, it actually was my browser. Well, I feel stupid for not checking that earlier. Thank you so much for your help. :)
[21:09:01] <GothAlice> javawolfpack: https://docs.mongodb.com/manual/tutorial/replace-replica-set-member/ < see also
[21:09:27] <GothAlice> Anon4254: No worries; I have to advise clients about broken browser extensions and stuff at work, too, since our sites actually use those headers. :)