mongodb process is stuck

[鐘士閔]

in these days , my mongodb process stuck very easy.

just like this mongostat cmd,

time from 07:19:00 jump to 07:24:00

between these two times,

I can't get any data from query cmd.

anyone knows what happen with mongodb?

[鐘士閔]

my mongodb version is 2.6 . 

鐘士閔於 2016年8月30日星期二 UTC+8下午4時05分16秒寫道：

[鐘士閔]

add more inforation

in that time, mongodb only can init connction , can't do anything else.

Please help me about this.

thank you very much

鐘士閔於 2016年8月30日星期二 UTC+8下午4時05分16秒寫道：

[鐘士閔]

mroe information  

Init Connect End,
always show this log.

 serverStatus was very slow: { after basic: 0, after asserts: 0, after backgroundFlushing: 0, after connections: 0, after cursors: 0, after dur: 0, after extra_info: 0, after globalLock: 0, after inde

xCounters: 0, after locks: 0, after network: 0, after opcounters: 0, after opcountersRepl: 0, after recordStats: 27682, at end: 27682 }

鐘士閔於 2016年8月30日星期二 UTC+8下午4時05分16秒寫道：

[Nishant Bhardwaj]

Hello,

Based on the mongostat output in your first post, I can see that there are lots of active reads. To start with I would recommend the following:

• Review Optimize Query Performance (v2.6) and Analyze Query Performance (v2.6) links in the MongoDB manual.
• Visualize database operations graphically using
  mplotqueries tool to plot queries within the specific time frames. i.e. 07:00:00 to 07:30:00. Mplotqueries’ scatter plot (the default) shows the duration of operations (y-axis) over time (x-axis) and makes it easy to spot long-running operations.
• Check your working set. This is the portion of your data that is accessed frequently, and should fit in memory to achieve good performance. See How can I estimate working set size (v2.6).

If you have further questions, could you provide the following relevant information:

• What is the configuration of your MongoDB setup? i.e. standalone, replica set, or sharded cluster.
• What is your underlying OS ? i.e. Windows, OSX, Linux Ubuntu, etc.
• What is the size of your working set and memory ?

You may also find the following resources useful:

• MongoDB Diagnostics FAQ (v2.6)
• MongoDB Optimization Strategies (v2.6)
• Blog: top 5 metrics to watch in MongoDB

Regards,

Nishant

​

[鐘士閔]

Hi Nishant ,

Thank you for your reply.

I am sorry about reply so late.

As you said 「a lot of active read」, do you mean 「query column」 or  「command column」?

I think that is because idle 5 min , so accumulate a lot of query or commands.

I will read your recommend document.

My MongodDB setup is Standalone,

OS：Cent OS7

Ram：60G 

When I use top command for linux stats , mongod  memory use about 40%,

VIRT about 76g

regards.

Nishant Bhardwaj於 2016年9月8日星期四 UTC+8下午10時45分50秒寫道：

[Nishant Bhardwaj]

Hello,

By active reads I meant queries, in the query column it is the number of query operations per second. You can read about mongostat in the mongodb manual.

Also please review MongoDB 2.6 production notes to ensure that your deployment follows the recommended settings for the best performance.

Regards,

Nishant

​

[鐘士閔]

Hi Nishant ,

We have done all recommended configuration that we can do on our server machine,

but still happen same issue.

We will try upgrade our mongodb version to 3.2 , this solution can help solve this issue ?

Nishant Bhardwaj於 2016年9月21日星期三 UTC+8下午3時03分58秒寫道：

[鐘士閔]

10/11 update

We have upgraded our mongodb from 2.6.1 to 3.2.9.

Now this issue not happen again.

鐘士閔於 2016年9月26日星期一 UTC+8下午2時36分31秒寫道：