cancel
Showing results for 
Search instead for 
Did you mean: 

SQL backup error code 2 - Error in curl async library., nbu status = 7643

hha_mea
Level 6
   VIP   

Greetings All,

We have a "strange" problem regarding SQL backup.

the job fails with error code 2 : none of the requested files were backed up

But in the logs we have certificate errors : 

nbpxyhelper:

27/06/2020 21:55:00.838 [ValidatePeerHostProtocol::handle_exception] Something serious happened with async library curl functions, will fail state machine (../machines/LibNbPxyValidatePeerHost.cpp:1205)
27/06/2020 21:55:00.838 [ValidatePeerHostProtocol::setState] Transitioning state from: VALIDATE_PEER_HOST_PROTOCOL_RUNNING to: VALIDATE_PEER_HOST_PROTOCOL_ERROR (../machines/LibNbPxyValidatePeerHost.cpp:1520)
27/06/2020 21:55:00.838 [ValidatePeerHostProtocol::setState] Error in curl async library., nbu status = 7643, severity = 2, Additional Message: [PROXY] Encountered error (VALIDATE_PEER_HOST_PROTOCOL_RUNNING) while processing(ValidatePeerHostProtocol)., nbu status = 1, severity = 1 (../machines/LibNbPxyValidatePeerHost.cpp:1539)

the strange thing is, when it fails, we re-run the job and it may work and it may not..

sometimes it works for 3-4 consecutive days and then it fails.

I have redeployed the certificate again (revoke+ reset host attributes) but still, the problem occurs.

these errors led me to this technote: 

https://www.veritas.com/support/en_US/article.100044566

and as our version is 8.2, we don't need an EEB because it is already included in the version (per the TN), so I just did the first & second steps and by doing that, the backup stopped working! (same error every time) so I guess we touched something in here !! I did a rollback and the backup worked again for 2-3 times then fails intermittently. 

my guessing then is that the problem comes from this but the solution provided by that technote isn't working.

Now, I am trying to install this EEB: 3998268_3 and I will do the 1-2 steps from the technote and cross fingers :)

so if anyone had this error before please advise :)

I will get back to you asap.

NBU version 8.2 (master, media & sql server)

 

Best regards.

 

 

1 ACCEPTED SOLUTION

Accepted Solutions

hha_mea
Level 6
   VIP   

I came back to mark this as a solution because the problem was resolved after discovering that the client 'sometimes' tries to reply back to the master server on the wrong IP address..

The master server has 2 IP addresses (one for production and one for backup) so when we try the bptestbpcd all is fine (even ping used to work..)  but sometimes it gets failing because the client chooses the wrong IP address to resolve the master's name. which explains why it works sometimes and also why we have certificate errors in the logs.

So the solution was to add the entry in the host's file on the client to force it to reply on the right ip address.

 

 

View solution in original post

7 REPLIES 7

hha_mea
Level 6
   VIP   

Any help guys? :)

Thanks,

Tape_Archived
Level 6
   VIP   

@HH can you verify if the client is validated with both fully qualified name & shortname? I see error 7643 in the message.

Please check under "Security Management => Host Management => Mappings for Approval" for this client. Using bptestbpcd make sure proper communication is established between client & master or media servers.

hha_mea
Level 6
   VIP   
Hello @Tape_Archived
Thank you for your reply,
Yes the error 7643 is all over the log and sometimes it work and sometimes no, and yes it is already mapped with short and full name I have also added the ip adress in case the resolution weren’t working.. still same problem.
As I said i believe the problem comes from the master as per the technote and also we must also install the EEB 3998268_3 and apply the 2steps from the tn (which is not done yet), I was wondering if anyone had the same problem..

Thiago_Ribeiro
VIP
Partner    VIP    Accredited

Hi @hha_mea ,

As you already created and recreated the certificate I think that problem can be the master server, but i cant confirm..

Did you already check this TN??

https://www.veritas.com/support/en_US/article.100043900

Hope this help.

Thiago

 

hha_mea
Level 6
   VIP   
Hello Thiago, thank you for your reply.
Yes I know this technote I actually applied that on other situations on which we had that problem but this one that we have there is no connection loss between the media and master server it concerns only this sql backup, and I don’t think neither that this is a certificate problem.. because As described before, it completly stopped working after applying that technote (step1 and 2) which is editing the web.conf file and activating WAL.. but I believe we still need to install the mentioned EEB because as soon as I apply those steps it stops working.. ( and unfortunately I still didn’t install that EEB) I will update this post as soon as I do it..
Many thanks again for your reply

hha_mea
Level 6
   VIP   
Besides, the problem in the technote you provided is now resolved in 8.2 (the version in our actual problem)

hha_mea
Level 6
   VIP   

I came back to mark this as a solution because the problem was resolved after discovering that the client 'sometimes' tries to reply back to the master server on the wrong IP address..

The master server has 2 IP addresses (one for production and one for backup) so when we try the bptestbpcd all is fine (even ping used to work..)  but sometimes it gets failing because the client chooses the wrong IP address to resolve the master's name. which explains why it works sometimes and also why we have certificate errors in the logs.

So the solution was to add the entry in the host's file on the client to force it to reply on the right ip address.

 

 

View solution in original post