Data Backup and Recovery

Question about snapmirror integration in vfiler setup

ANDYOEAMTC
3,831 Views

Hi!

I'm testing snapcreator with our MaxDB installations. So far things are functional, but if I activate snapmirror integration I got an error:

ZAPI: Snapmirror error: source specified does not match configuration file

ERROR: [scf-00013] NetApp Snapmirror Update on destination filer2:vol1 failed! Exiting

Our environment looks like:

physical filer1 with vfiler1 has vol1, vol2 and vol3.

physical filer2 is destination for snapmirror of mentioned volumes.

As controller I use filer1 because we can't make snapshots, a.s.o. via vfiler context...

Snapmirrors are configured with vfiler1 as source because of better useability - but it seems there is the problem, or am I wrong?

Any hint how we can fix this?

Thanks in advance!

Andy

7 REPLIES 7

ktenzer
3,831 Views

If the snapmirror relationships are inside the vfiler then you need to specify the vfiler in the VOLUMES parameters.

SC does not have ability to tunnel through physical filer to vfiler. We need to connect directly to physical or vfiler.

Change following parameters:

TRANSPORT=HTTP

VOLUMES=vfiler1:vol1,vol2,vol3

SNAPMIRROR_VOLUMES=vfiler1:vol1,vol2,vol3

NTAP_USERS=vfiler1:user/pass

Regards,

Keith

ANDYOEAMTC
3,831 Views

Hi!

First of all thanks for your rapid answer!

If I change every "filer1" to "vfiler1" then I get the following error:

...

########## PRE NETAPP COMMANDS FINISHED SUCCESSFULLY ##########

<system-get-ontapi-version></system-get-ontapi-version>

[<timestamp>] TRACE: ZAPI RESULT

<result status="failed" reason="in Zapi::invoke, cannot connect to socket" errno="13001"></results>

[<timestamp> ERROR: [scf-00013] in Zapi::invoke, cannot connect to socket at /<path-to-snapcreator-server>...

...?

If I leave all entries on filer1 and change only the SNAPMIRROR_VOLUMES line to vfiler1:vol1,vol2,vol3, then backup is running through, but snapmirror is ignored...

Any ideas?

Thanks,

Andy

ktenzer
3,831 Views

Hi Andy,

Yep this error is network related. Here is what it could be:

1. vfiler1 has no DNS or hosts entry so host resolution fails. A "ping vfiler1" from scServer would tell you if that is problem.

2. TRANSPORT=HTTP must be set. A vfiler only supports HTTP for communication. A "telnet vfiler1 80" from scServer will tell you if that is problem.

3. Firewall could be issue so you may be blocking communication to vfiler

4. Maybe http isnt enabled on vfiler1? An "options http" should show you. The http.admin.enable must be on, this is what allows HTTP connection.

As for chaning SNAPMIRROR_VOLUMES to vfiler1 when VOLUMES is set to filer1, yes in this case snapmirror is ignored since no filer in SNAPMIRROR_VOLUMES matches what you have in VOLUMES.

Regards,

Keith

ANDYOEAMTC
3,831 Views

Hi!

On vfiler-context we have no httpd.admin - options:

vfiler1:

#> options |  grep http
httpd.access                legacy    
httpd.autoindex.enable      off       
httpd.bypass_traverse_checking off       
httpd.enable                off       
httpd.log.format            common    
httpd.method.trace.enable   off       
httpd.rootdir               XXX       


phys. filer1:

#> options |  grep http
httpd.access                legacy    
httpd.admin.access          legacy    
httpd.admin.enable          on        
httpd.admin.hostsequiv.enable off       
httpd.admin.max_connections  512      
httpd.admin.ssl.enable      off       
httpd.admin.top-page.authentication  on        
httpd.autoindex.enable      off       
httpd.bypass_traverse_checking off       
httpd.enable                off       
httpd.log.format            common     (value might be overwritten in takeover)
httpd.method.trace.enable   off       
httpd.rootdir               XXX       
httpd.timeout               300        (value might be overwritten in takeover)
httpd.timewait.enable       off        (value might be overwritten in takeover)

Is this due to our ontap version (8.0.1p3)?

Thanks & bye,

Andy

ktenzer
3,831 Views

Ok not sure could be mistaken on if vfiler have http.admin.enable

Anyway if you can "telnet vfiler1 80" then it is working if not you need to allow access

Keith

ANDYOEAMTC
3,831 Views

Hi!

It seems that in our onTap-Version there is no way to activate http on vfiler...

I've played with some parameters:

When setting the following:

VOLUMES=filer1:vol1,vol2,vol3

SNAPMIRROR_VOLUMES=filer1:vol1,vol2,vol3

SECONDARY_INTERFACES=filer1:vfiler1/filer2

log:

...

########## Running NetApp Snapmirror Update using source filer1:vol1 ##########

[<timestamp>] INFO: Getting NetApp Snapmirror Status for filer1

[<timestamp>] INFO: NetApp Snapmirror Status for robin finished successfully

[<timestamp>] INFO: Secondary Interfaces Detected! Source vfiler1 and Destination filer2 will be used for snapmirror

[<timestamp>] INFO: Secondary Interfaces Detected! Source vfiler1 and Destination filer2 will be used for snapmirror

########## Detecting Data OnTap mode for filer2 ##########

[<timestamp>] INFO: Data OnTap 7 mode detected

[<timestamp>] INFO: NetApp ASUP create on filer2 finished successfully

[<timestamp>] INFO: Running NetApp Snapmirror Update on destination filer2:vol1 source vfiler1:vol1

[<timestamp>] ZAPI: Snapmirror error: source specified does not match configuration file

[<timestamp>] ERROR: [scf-00013] NetApp Snapmirror Update on destination filer2:vol1 failed! Exiting

...

So it looks good as it now says source vfiler1:vol1 and destination filer2:vol1 - but ...

Any ideas?

Thanks & bye,

Himmler Andreas

ktenzer
3,831 Views

You need to check http options on filer2

cuba2> options http

httpd.access                 legacy    

httpd.admin.access           legacy    

httpd.admin.enable           on        

httpd.admin.hostsequiv.enable off       

httpd.admin.max_connections  512       

httpd.admin.ssl.enable       on        

httpd.admin.top-page.authentication on        

httpd.autoindex.enable       off       

httpd.bypass_traverse_checking off       

httpd.enable                 on        

httpd.log.format             common     (value might be overwritten in takeover)

httpd.method.trace.enable    off       

httpd.rootdir                /vol/vol0/home/http

httpd.timeout                300        (value might be overwritten in takeover)

httpd.timewait.enable        off        (value might be overwritten in takeover)

Make sure that http.admin.enable is on and if you are doing HTTPS you need http.admin.ssl.enable

Regards,

Basically what happens when you do "telnet filer2 80" from Snap Creator server? Do you get a connection? If not why not? DNS / Hosts resolutions, firewall, configuration settings on storage system?

Keith

Public