App crash due to non existent id

[copleykj]

I have an app running on modulus. Everything ran relatively smoothly until I scaled it to a second servo and now all the sudden I'm getting this error and then the app instance crashes..

Error: 53fcdaea80eba4530628442f does not exist
    at /mnt/data/1/node_modules/mongodb/lib/mongodb/gridfs/gridstore.js:205:24
    at /mnt/data/1/node_modules/mongodb/lib/mongodb/cursor.js:744:35
    at Cursor.close (/mnt/data/1/node_modules/mongodb/lib/mongodb/cursor.js:969:5)
    at Cursor.nextObject (/mnt/data/1/node_modules/mongodb/lib/mongodb/cursor.js:744:17)
    at commandHandler (/mnt/data/1/node_modules/mongodb/lib/mongodb/cursor.js:713:14)
    at /mnt/data/1/node_modules/mongodb/lib/mongodb/db.js:1843:9
    at Server.Base._callHandler (/mnt/data/1/node_modules/mongodb/lib/mongodb/connection/base.js:445:41)
    at /mnt/data/1/node_modules/mongodb/lib/mongodb/connection/server.js:468:18
    at MongoReply.parseBody (/mnt/data/1/node_modules/mongodb/lib/mongodb/responses/mongo_reply.js:68:5)
    at null.<anonymous> (/mnt/data/1/node_modules/mongodb/lib/mongodb/connection/server.js:426:20)

[aldeed]

Scaling on modulus doesn't somehow shard the database, does it? That's all I can think of.

I don't think we've ever tested with sharded gridfs.

http://docs.mongodb.org/manual/tutorial/shard-gridfs-data/

[copleykj]

The database is actually on compose's (formerly mongoHQ) elastic deployment which is a three member replica set. I would assume that scaling on modulus wouldn't affect this. Fortunately for the moment I've been able to dial it back to just a single servo but would be great to figure it out before I have to bump it up again.

[aldeed]

cc @vsivsi in case he has any ideas

[justinventuri]

We have this issue, too. Usually the app restarts without incident. However, there are times when we need to go into the db and delete all temp collections. We are using gridfs and s3.

[jimmiebtlr]

Was anything additional determined on this? I'm also experiencing this in current versions.

[copleykj]

I would love to say that I figured this out, but unfortunately this issue caused me to move to using cloudinary.

[mxab]

do you guys use multiple nodes? As far as I know this won't work, but you can try out @rhyslbw pull request #667

[neopostmodern]

The duplicate bug #731 happened on my local development machine too, definitely a single node.

[ClarenceL]

Getting the same thing, Error: 55df664eebec6253187bd231 does not exist

What id is this referring to?

[derwaldgeist]

Had the same error when trying to upload more than one (server) local file to GridFS in one batch.

[MikkoA]

Same error on my local dev machine, crashes Meteor. But after Meteor restarts, browser tries to send PUT request again and the file is uploaded properly. Any ideas on this one?

[ClarenceL]

Seems to be GridFS, use cfs:filesystem.. don't remove the GridFS package though, it seems it's still needed if you use mup to deploy

[vietnv22]

+1