So I am trying to restart mongodb 3.2 and enable authorization, so i edit the /etc/mongod.conf file and added
security:
authorization:enabled
then i saved the file, and typed in
sudo service mongod restart
which showed that it restarted correctly, but when I looked at the processes running, mongod is not one of them.
And now I cant restart it at all.
Also, there was already a database with information in monogodb before i enabled authorization. Im not sure if that is important to know.
I checked out the solutions in here Stop: Unknown instance mongodb (Ubuntu)
but i dont have a " fork = true" statement anywhere in /etc/mongod.conf
Okay, so strange enough, I did the following
1) went back to the /etc/mongod.conf file and commented out the authorization configuration.
2) typed in sudo service mongod start and gave me back the the job is already running.
3) checked the the running processes and mongodb was now running properly! I even accessed it through the monoShell
4) then i stopped the service sudo service mongod stop
5) I went back and add the following exact syntax
security
authorization: enabled
6) saved the file and sudo service mongod start
and waaar'yaaa know. its working.....
can someone explain ? that would be helpful
Related
I unable to start MonogoDB service after adding users into admin db as well as my db.
I setup MonogoDB and started service using following command i.e.
service mongod start
Using command prompt, I added few users like dbOwner, clusterAdmin, readWrite, read roles base users. Along with that I also changed configuration from /etc/mongod.conf. in that file, I changed port number, IP addresses, dbPath, and security.authorization: enabled.
Then I restarted mongod service using following command.
service mongod restart
After ran this command, mongod service stopped successfully, but failed to start with only 'FAILED' message.
I tried execute following command i.e.
mongod --port 27123 --dbpath /path/to/db --auth
It is working.
Question: How to execute 'service mongod start' using additional parameters in CentOS?
MonogoDB: 3.4
OS: CentOS 7
I got solution i.e.
mongod --config /etc/mongod.conf
Referred: https://docs.mongodb.com/manual/reference/configuration-options/#use-the-configuration-file
It starts child process and also I can stop mongod service using service mongod stop command.
But I don't know whether it is correct or not.
I can't certify exactly where the script that "service" command uses on CentOS 7, but in Ubuntu 18.04 mongod service script file is in
/lib/systemd/system/mongod.service
There you can change the user who executes the process and add any parameters you want, like --auth.
Said that, if you ever executed mongod as root, some files on where you store the db data will have the owner as root, making the database fail to start as another user. The fix I found for that is to manually chown to mongodb:mongodb (or the user you want to use) all the files that are owned by root inside the database.
Hope this helps.
mongod.service file from mongodb github
I am using Ubuntu 14.04 LTS and I have installed MongoDB 3.2.1. I had various problems with it that I fixed by either finding help from the internet (thanks Google) or by uninstalling and re-installing MongoDb.
One persistent problem that I cannot fix, unless I re-install, is by running mongod on the terminal. Currently my mongodb installation is working properly. With sudo service mongod start I start mongodb (I commented out start on xxxx line at /etc/init/mongod.conf so it doesn't auto start.) and with sudo service mongod stop I can stop it normally atm. And it correctly uses /var/lib/mongodb path for saving my collections.
If however after I stop mongodb with sudo service mongod stop I hit, on the terminal, mongod then mongodb breaks. I have gotten all kinds of errors like
Unusable mongod.lock. I have gotten around it, in a previous install, by doing sudo mongod but that was not a proper solution.
/data/db folder not found !!. Why look for it in the first place? The /etc/mongod.conf specifies the dbpath /var/lib/mongodb which is also the default when mongodb gets installed !!?? I have gotten around it as well with the mongod --dbpath /var/lib/mongodb option.
I think there was one more error that I don't remember but I also fixed/gotten_around it by finding solutions online.
I know that by re-installing and by never running mongod I can - for now? - not face those problems again (which looks a much better solution that the "workarounds" I did whenever those problems arose. I am wondering however what it is that is going wrong when I type mongod instead of sudo service mongod start ?!
In case somebody tries to replicate the problem know that I ve done only two modifications on my system after installing:
Commenting out the start on xxxx line at /etc/init/mongod.conf
Disabling transparent hugepages as described on the answer here.
I don't think anyone of those should interfere with my installation.
Can anyone help me understand what is going on? Aren't those commands supposed to do the same thing??
Thanks for your help.
Ok, I will try to answer.
First, unless you know what you are doing, you should not start mongod manually.
In general (a bit simplified), calling [sudo] service mongod start, you instruct the system to read the according file in /etc/init and start the executable according to the configuration described in said file.
When you started mongod by hand, however, you actually called the mongod binary, the server software itself – while the name is the same as the service, the two commands have few things in common. The binary does not use the /etc/mongod.conf by default, falling back to its default values for the various settings. Actually, you can see that the config file is explicitly defined in the init script. This is why mongod tried to find /data/db.
You can find said binary by issuing
which mongod
Regarding the lock file: When mongod is started by the system, user root actually assumes the effective user id of mongod (or mongodb I don't know for Ubuntu of the top of my head). When you tried to start it from your user id, you do not have the privileges to overwrite the lock file. When you used sudo mongod afterwards, you assumed the effective user id of root which on the other hand is allowed to overwrite said file. However, mongod will the run as root which is a security no-no.
An init script defines which environment to use, which user to run under and a lot of other stuff. Unless you really know what you are doing, you should not even fiddle with them, much less skip it.
And now, with the finger up
sudo is not the UNIX way of saying "I mean it!"
It has security implications, and you should be very aware of those implications before using it – aka read the man pages of every command you issue before you use it until you have at least a fact based idea of what the command is doing.
And again: unless you really know (as opposed to assume) what you are doing, do not fiddle with the system configuration.
I have mongodb installed on a new Amazon Linux AMI after implementing the guide here. When running "service mongod start" though, I just get a Starting mongod FAILED message. Nothing else. Blank log file also.
service mongod stop yields a FAILED also.
service mongod status yields "mongod is stopped"
Any thoughts or next steps?
I faced the same problem with my master mongodb instance. I have a replica set of 3 instances on EC2 and I had the same problem after one reboot that we needed for it, mongod simply failed to start.
First thing I did, was changed the ownership of the log file that is set up in the mongod.conf:
sudo chown mongod:mongod <LOG FILE>
After this I tried only sudo mongod and this will log the problem why the service won't start in the log file.
Mine appeared to be some ownership issues with the WiredTiger conf files in the dbPath directory (were set to root ownership instead to mongod user)
What I do in that case is use
sudo killall mongod
and after that it starts ok. My suspicion is that there is a running instance and that's why it fails to start.
It has been a while since I have used Mongodb and I am running into a problem. I have an EC2 instance that had Mongodb running on it, but I had turned off the instance for a few months. After turning the instance back on, Mongodb will not start correctly. No log file is being created so I do not know what is causing the problem.
When I type:
sudo service mongodb start
I get the message:
mongodb start/running, process 2432
but mongo is not actually running and a mongod.lock file is being created.
If I type:
mongod
by itself, Mongodb will actually start and be usable. But, I must leave the terminal window up and not cancel out of the command to keep it running. I have to open up a second terminal window to access my databases. I guess my main question is how I get the mongod service to stay running itself without having to leave a terminal window open with the command running.
Sorry if my explanation doesn't make sense. Hopefully you get what I mean. Anyone have any idea what I am missing? I had this problem before months ago and was able to solve it. Sadly, I didn't write down the missing ingredient.
For that, run command using Nohup
nohup mongod &
Then you can terminate you terminal.
I am new to Mongodb. When I run the following command:
sudo service mongodb start
It shows me "mongodb start/running, process 3566". But when I try to stop by giving the below command:
sudo service mongodb stop
It shows stop: Unknown instance:. Do anybody have any idea how to slove this issue.
And if I give sudo service mongodb restart, I'm getting the below message:
stop: Unknown instance:
mongodb start/running, process 3644
PS: If I'm using just mongod, I could start the server and could connect the mongo shell by giving mongo.
Please suggest me how to fix the issue and how to connect mongo shell after giving service mongodb start. Thanks in advance.
I just ran into this.
Chances are the issue is in your conf file (obviously, since you reinstalled the seeded)
In the mongodb.conf do not set fork = true if calling it as a service. This will prevent the service call from being able to access it with service mongodb.conf status/stop/restart.
Removing a fork=True line, however, will alleviate the issue.
See comments; use the other solution even though this one is accepted.
The reason you get the Unknown instance error is because upstart is tracking the wrong PID for the instance of forked mongod process. But the proper solution to this is not running the process in foreground by removing fork=true line from the config like #GoingTharn suggesting, but rather helping upstart to capture the right PID of the forked process by adding
expect daemon
to the /etc/init/mongodb.conf
I faced exactly the same problem and did the followings:
Removed the file mongod.lock (in my case this file was located at /var/lib/mongodb/)
Used the following command from my mongo client
mongo --repair
And that fixed the issue. It may be noted that I had a system crash prior to this problem appeared. Therefore, what I believe, the problem was caused due to the unclean shutdown of the server demon.
If you are unsure of the reason, you may wish to have a look at the log file (in my case I found it under /var/log/mongodb/). That might give you some useful hints.
Thanks.
I faced it as you had.Maybe it is because my computer crash unexpectedly.I use the command: sudo top(ubuntu), and I found mongo running.I killed it and start mongo.Every thing turns right.
I fixed it by myself. Below is what I did:
apt-get install mongodb
And overwrote .conf file. After that everything works fine. I hope this would help someone.