SP_connect: DEBUG: Auth list is: NULL

Receiving this message when I try and restart one of the nodes in a three node cluster.  That node is unable to restart:

 

/opt/vertica/bin/vertica --status -D /vjanys/dbadmin/vjanys/v_vjanys_node0002_catalog
vertica process is not running

 

spread is running:

 

 /opt/vertica/spread/sbin/spread -n N010000000104 -c /opt/vertica/config/vspread.conf

Comments

  • SruthiASruthiA Administrator

    Hi,

     

     

       Can you check if SSL is enabled?

     

     

    -Regards,

     Sruthi

  • Yes, SSL is enabled in /vjanys/dbadmin/vjanys/v_vjanys_node0002_catalog/vertica.conf:

     

        EnableSSL = 1

     

  • SruthiASruthiA Administrator

    Hi,

     

      please edit the vertica.conf file to disable SSL. (EnableSSL = 0 will disable it). Please do it on all the nodes and try restarting the database. Let me know how it goes.

     

     

    -Regards,

     Sruthi

  • I set  EnableSSL = 0 on all hosts and then "stopped vertica on host" (through the admintool) for each host, then again started the db (through the admintool); the vertica process on that node will still not start:

     

    Node Status: v_vjanys_node0001: (UP) v_vjanys_node0002: (DOWN) v_vjanys_node0003: (UP)

     

  • SruthiASruthiA Administrator

    Hi,

     

      Can you share me the vertica.log file

     

    -Regards,

     Sruthi

  • SruthiASruthiA Administrator

    Hi,

     

       Try force restarting the node2. I think that may help.

     

    /opt/vertica/bin/admintools -t restart_node -s <this_Hostname_or_IP> -d dbname -p password -F

     

    -Regards,

     Sruthi

  • Yes, this worked.  Thanks!

  • It has been in RECOVERY mode for a few days now.  Is this normal?

     

      │  vjanys_| 10.0.0.104_| RECOVERING_    │  
      │  vjanys_| 10.0.0.110_| UP_________    │  
      │  vjanys_| 10.0.0.140_| UP_________ 

  • Here is the line for recovery_status:

     

     v_vjanys_node0002 |        193827 | historical pass 1 |                0 |            0 |                   12 |               44 |                 0 |             0 | t

  • After doing a bit of research, it seems that a recovery taking days and/or weeks is possible.

  • After doing a bit of research, it seems that a recovery taking days and/or weeks is possible.

     

    Long recoveries are possible, but in general recovery should be pretty quick.  And you want it to be quick so that HA is restored as quickly as possible.  

     

    If your recovery is slow, you should check why.  Is this one long recovery or recovery failing and retrying?

     

    - does the PROJECTION_RECOVERIES table show anything interesting?

     

    - does the vertica.log on the recovering node show any errors?

     

      --Sharon

     

Leave a Comment

BoldItalicStrikethroughOrdered listUnordered list
Emoji
Image
Align leftAlign centerAlign rightToggle HTML viewToggle full pageToggle lights
Drop image/file