[SERVER-25299] shut down Created: 27/Jul/16 Updated: 24/Aug/16 Resolved: 24/Aug/16 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Replication |
| Affects Version/s: | 3.2.5 |
| Fix Version/s: | None |
| Type: | Bug | Priority: | Critical - P2 |
| Reporter: | Ian Sharp [X] | Assignee: | Eric Milkie |
| Resolution: | Done | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Environment: |
linux ec2 |
||
| Issue Links: |
|
||||||||
| Operating System: | ALL | ||||||||
| Steps To Reproduce: | make 3 mongo instances (each on their own EC2 linux amazon AMI), behind an AWS load-balancer ELB using HTTPS and Amazon's SSL certs |
||||||||
| Participants: | |||||||||
| Description |
|
a mongoDB instance shutdown while running as part of a replica set on an Amazon EC2 AMI instance uname -a
df -h
mongo.log
|
| Comments |
| Comment by Eric Milkie [ 08/Aug/16 ] |
|
Hi Ian, |
| Comment by Ian Sharp [X] [ 30/Jul/16 ] |
|
Thanks Eric Milkie, any ideas why the queries were taking 17 seconds, instead of 10's of milliseconds ? i noticed some of the queries with several embedded fields taking a few seconds (but not 17). Or could it possibly be the fact that sometimes i open a database connection, then do a gauntlet of analytics on the data (taking sometimes up to 30 seconds) and then close the database connection? (probably doesn't matter since the database response still comes quickly? or does leaving the database connection open while analytics run, give the risk of replica set failovers and loss of service?) ... for example, i may have a function that leads to 6 or so analytics callbacks, and i pass the database object onto each successive analytics function, so that i can keep querying the replicaset without 'closing & re-connecting to mongodb replica set again'. I thought this would be efficient, but perhaps this keeps the database object alive too long and i'd be better off getting a new database object by closing the previous connection and connecting to mongo again? |
| Comment by Eric Milkie [ 29/Jul/16 ] |
|
Hi Ian, |
| Comment by Ian Sharp [X] [ 29/Jul/16 ] |
|
This is how i start each instance sudo mongod --fork --port 27017 --replSet go --logpath ~/log/mongo.log |
| Comment by Ian Sharp [X] [ 29/Jul/16 ] |
|
Andy Schwerin, I uploaded: instance1_rs_conf.txt |
| Comment by Andy Schwerin [ 28/Jul/16 ] |
|
DrIanSharp@gmail.com, can you share the replica set configuration for your replica set? I believe you can use rs.conf() in the shell to get it. |
| Comment by Ian Sharp [X] [ 28/Jul/16 ] |
|
idk if I got low on RAM? free -m |
| Comment by Ian Sharp [X] [ 28/Jul/16 ] |
|
i'm also pretty sure i followed the directions here correctly: https://docs.mongodb.com/ecosystem/platforms/amazon-ec2/ which helped mongo from shutting down in the past |
| Comment by Ian Sharp [X] [ 27/Jul/16 ] |
|
Thanks Thomas Schubert. I uploaded mongo.log it's about 155MB |
| Comment by Kelsey Schubert [ 27/Jul/16 ] |
|
I have created a secure upload portal for you to use. Files uploaded to this portal are only visible to MongoDB employees investigating this issue and are routinely deleted after some time. Would you please upload the complete logs to this portal? Thank you, |
| Comment by Ian Sharp [X] [ 27/Jul/16 ] |
|
Hi Andy Schwerin, there is no more log after the end of backtrace. there is more log before it. |
| Comment by Andy Schwerin [ 27/Jul/16 ] |
|
Could you attach about 50 more log lines above what you posted, or the complete logs if you have them available? |