[DOCS-13847] [OM] Fix Errors in 4.2 Manual Restore Instructions Created: 24/Aug/20  Updated: 29/Oct/23  Resolved: 11/Jan/21

Status: Closed
Project: Documentation
Component/s: Ops Manager
Affects Version/s: None
Fix Version/s: None

Type: Task Priority: Major - P3
Reporter: Greg Cox Assignee: Julia Malkin
Resolution: Fixed Votes: 0
Labels: new-hire-ticket
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Issue Links:
Backports
Duplicate
Problem/Incident
Related
Participants:
Days since reply: 3 years, 3 weeks, 1 day ago
Epic Link: DOCSP-3127
Story Points: 5

 Description   

Description

https://docs.opsmanager.mongodb.com/v4.2/tutorial/restore-replica-set/

There are errors with the instructions in the manual restore instructions for a Replica Set as per link above:

FCV of 4.2 or later > Manual Restore

  1. Step 7 - should state what do with existing files - delete them ? 
  2. Step 10 - doesn't state what user to use and what permissions are required
  3. Step 13 - Errors with restoreTS.getTime() not recognised. This there is a missing declaration of restoreTS
  4. Step 13 - My  restoreInfo.txt did not contain a "Restore timestamp"
  5. Step 14 - Refers to a port number in Step 10 that doesn't exist
  6. Step 17 refers to connecting to a mongo shell but we were instructed to shut this down in Step 15. Step 16 was skipped because it states to do so if not a PiT restore so maybe a mistake here ?

Scope of changes

Impact to Other Docs

MVP (Work and Date)

Resources (Scope or Design Docs, Invision, etc.)



 Comments   
Comment by Julia Malkin [ 19/Jan/21 ]

gaurav.guptaThank you for taking a look! I have discussed this issue with Giordano and Steve. Here is what I did as a result:

  • Transferred some items that are applicable to both topics to the Sharded Clusters procedure.
  • Did not overhaul the Sharded Clusters procedure overall for these reasons: it differs from the RS procedure enough to warrant a separate ticket; it is out of scope for this ticket; and, most importantly, it was not clear from this ticket what are the issues with the Sharded Clussters procedure specifically. If you know that customers are having issues with Sharded Clusters procedure, and have specific items in it that must be addressed, please lot a ticket and assign to me. I will look into it separately. Thank you again for taking care of our docs and for you review!
Comment by Gaurav Gupta [ 19/Jan/21 ]

julia.malkin Seems like the update to doc was not replicated to sharded cluster manual PIT Restore.

https://docs.opsmanager.mongodb.com/current/tutorial/restore-sharded-cluster/

Comment by Githook User [ 11/Jan/21 ]

Author:

{'name': 'JuliaMongo', 'email': '72459727+JuliaMongo@users.noreply.github.com', 'username': 'JuliaMongo'}

Message: (DOCS-13847) Fix Errors in 4.2 Manual Restore Instructions (#3514) (#3525)

  • (DOCS-13847) Fix Errors in 4.2 Manual Restore Instructions
  • (DOCS-13847) Fix Errors in 4.2 Manual Restore Instructions-add Giordano's feedback
  • Included Giordano's and Steve's review comments.
  • Added Steve's corrections
  • Round three: Untangling mongod and mongo shell commands. Steve's review.
  • Getting ready for tech review
  • Include tech review comments from Steven and Giordano
Comment by Julia Malkin [ 11/Jan/21 ]

Addressed all review comments, transferred some applicable changes to the Sharded Clusters manual restore procedure (that is similar in some ways), merged, and closed.

Comment by Githook User [ 11/Jan/21 ]

Author:

{'name': 'JuliaMongo', 'email': '72459727+JuliaMongo@users.noreply.github.com', 'username': 'JuliaMongo'}

Message: (DOCS-13847) Fix Errors in 4.2 Manual Restore Instructions (#3514)

  • (DOCS-13847) Fix Errors in 4.2 Manual Restore Instructions
  • (DOCS-13847) Fix Errors in 4.2 Manual Restore Instructions-add Giordano's feedback
  • Included Giordano's and Steve's review comments.
  • Added Steve's corrections
  • Round three: Untangling mongod and mongo shell commands. Steve's review.
  • Getting ready for tech review
  • Include tech review comments from Steven and Giordano
Comment by Julia Malkin [ 05/Jan/21 ]

ulrich.blumensaatI have addressed your comments and many many other comments. The doc is now in tech review.

Comment by Julia Malkin [ 19/Dec/20 ]

The pull requests that are merged are for another ticket. I messed up. This ticket does not yet have pull requests.

Comment by Gaurav Gupta [ 04/Dec/20 ]

Through HELP-20278, we were able to prepare working steps for the manual PIT restore of 4.2 backups.
Please check the case which can help in correcting the documentation .

Thanks steven.connors!

Comment by Gaurav Gupta [ 25/Nov/20 ]

I have opened HELP-20278 for this issue as PIT Manual restore is still failing while applying oplogs from MBRU

Comment by Gaurav Gupta [ 26/Oct/20 ]

I also have a case on which customer is having trouble following these manual restore steps for OM 4.2

https://support.mongodb.com/case/00712525

Comment by Giordano Fusco (Inactive) [ 28/Aug/20 ]

I tried step by step a manual restore. These are the updates steps for both Cloud and OM, both 4.2 and 4.4:

  • 1 to 6 ok
  • new 7) shut down destination replica set
  • 8 and 9 ok
  • new 10) content of old 7
  • new 11) start as a standalone:

    Issue the following command:
    ./mongod --dbpath </path/to/datafiles> \
             --port <ephemeralPort> \
    

    Since this is the first time we use mongod, please add a note like to old 17: Depending on your path, you may need to specify the path to the mongod.

  • new 12) like old 17, but with <ephemeralPort> instead of just <port>
  • new 13) old 10
  • new 14) like old 11, but with <ephemeralPort> in the 2 places where it has <port>
  • new 15) old 12
  • new 16) like old 13, but add that restoreTS is the Restore timestamp from the restoreInfo.txt file.
  • new 17) Shut down the standalone
    Issue the following command:

    db.getSiblingDB('admin').shutdownServer()
    

  • new 18) like old 14, but without the important box
  • new 19) Shut down the standalone

    ./mongo --port <ephemeralPort> --eval "db.getSiblingDB('admin').shutdownServer()"
    

  • new 20) old 16
  • new 21) like old 14, but without the important box
  • new 22) like old 17, but with <ephemeralPort>
  • new 23) like old 18, but adding quotes and more details:

    rs.initiate( {
      _id : "<replaceMeWithTheReplicaSetName>",
      "version" : NumberInt(1),
      "protocolVersion" : NumberInt(1),
      "members" : [
        {
          "_id" : <id>,
          "host" : "<host>:<port>"
        },
       <allOtherMembers>
      ]
    })
    

  • new 24) Shut down the new Replica Set.

    db.getSiblingDB('admin').shutdownServer()
    

  • new 25) Repeat Steps 10 to 24 for each node in the Replica Set.
  • new 26) like old 21
Comment by Anthony Sansone (Inactive) [ 26/Aug/20 ]

greg.cox: points 3 & 4 in this issue are potentially a bug. Could you file a CLOUDP ticket to engage engineering? giordano.fusco had explained that the Restore timestamp should always be in that file.

Comment by Greg Cox [ 26/Aug/20 ]

Hi Anthony. No it didn't. I attempted the instructions the other day and ran into these issues

Comment by Anthony Sansone (Inactive) [ 26/Aug/20 ]

greg.cox: Did this surface from a HELP ticket? There may be a code fix that's required.

Comment by Anthony Sansone (Inactive) [ 25/Aug/20 ]

giordano.fusco: We worked on this in DOCSP-7114 (See PR #2779 to refresh your memory.)

Do these changes look appropriate?

Generated at Thu Feb 08 08:08:52 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.