San client not discovering fibre transport devices

[ldias]

I'm trying to install a san client. But from NBU admin console, on  SAN Client section I can't see any fiber transport device found.

I followed Netbackup San Client Installation and Configurarion Guide. Both installation (FT media server and san client ran ok without errors)

Some informations:

-FT Media Server is a Linux RHEL6.3 64bits running NBU 7.5.0.4

-San Client is a RHEL6.2 64 bits running NBU Client 7.5.0.4

-Checked san client licence on master server.

-SAN zoning configured properly

-FT Media Server HBA Qlogic ok

-From san client set 'bpclntcmd -sanclient 1

-From NBU administration console

-From san client operating system I can see two device "ARCHIVE  Model: Python" (cat /proc/scsi/scsi). So physically it sees device from ft pipe.

-Sending screenshots from NBU Administration Console.

Need help.

[Nicolai]

Has the nbftclient been re-started after the "Archive Python" devices has been zoned into the host ?

Two handy command as well:

nbftconfig -listclients -verbose

nbftconfig -listtargets -verbose

[ldias]

Yes, we have re-started san client services after the "Archive Python".

The output of nbftconfig -listclients -verbose shows me the client but FT Server Connections is 0:

"SAN Client Name      : yyyyyyyyy
SAN Client Version   : 7.5
SAN Client State     : active
Master Server Name   : zzzzzzzzz
FT Server Connections: 0
Client Ports/Server  : 2
Usage Preference     : preferred
Backup Wait Time     : 15
Restore Wait Time    : 5"

The output of nbftconfig -listtargets -verbose shows me that media server ft has pipes available.
FT Server Name   : xxxxxxxx

FT Server HBA Port   : 0
FT Server Port WWN   : 21:00:00:XX:XX:XX:XX:XX
FT Server Port Mode  : FABRIC
FT Server Port Model : QLE246x Series FC Hba
FT Server Port Vendor: Qlogic
        FT Server Device State   : active
        Associated LUN           : 0
        Active Connections on LUN: 0

        FT Server Device State   : active
        Associated LUN           : 1
        Active Connections on LUN: 0

It seems something related to registration of san client on ft media server.

Any ideas ?

[huanglao2002]

Please reference the attachment.

[ldias]

Can't see any logs from client (nbftclnt logs)

vxlogview -p 51216 -o 200 -t 24:00:00 shows nothing.

I already set logs DebugLevel=6

On ft media server side I got logs from nbftsrv

05/22/2013 17:27:45.462 [TestPipe] Sending VRTS_ASCQ_DATA_NOT_READY check condition to PN:IID:TID:LUN[1:1:0:1]
05/22/2013 17:27:45.462 [TestPipe] Sending VRTS_ASCQ_DATA_NOT_READY check condition to PN:IID:TID:LUN[1:1:0:0]
05/22/2013 17:37:03.668 [ProcessManage] node IOTID[1,1,0,65535] 0x7f2560008880 STMTMTYPE_PortDBChange type 0 STMTMLoggedIn
05/22/2013 17:38:12.392 [TestPipe] Sending VRTS_ASCQ_DATA_NOT_READY check condition to PN:IID:TID:LUN[1:1:0:1]
05/22/2013 17:38:12.392 [TestPipe] Sending VRTS_ASCQ_DATA_NOT_READY check condition to PN:IID:TID:LUN[1:1:0:0]
05/23/2013 08:49:19.493 [TestPipe] Sending VRTS_ASCQ_DATA_NOT_READY check condition to PN:IID:TID:LUN[1:1:0:1]
05/23/2013 08:49:19.496 [TestPipe] Sending VRTS_ASCQ_DATA_NOT_READY check condition to PN:IID:TID:LUN[1:1:0:0]
05/23/2013 15:32:39.048 [TestPipe] Sending VRTS_ASCQ_DATA_NOT_READY check condition to PN:IID:TID:LUN[1:1:0:1]
05/23/2013 15:32:39.048 [TestPipe] Sending VRTS_ASCQ_DATA_NOT_READY check condition to PN:II

Any ideas ?

[Juannillus]

Hello Luiz,

could you send the output of bpps -a from the client?

I had found in some machines in the past that to be able to see the FT devices I had to restart NBU services in the client including PBX. It doesn't happen all the time, but it has happened to me in several machines.

Send the output from the client command and try to restart all NBU services in client including PBX and tell us how it goes.

Regards,

Juan

[ldias]

Here goes the output of bpps -x

NB Processes
------------
root     23460     1  0 10:13 ?        00:00:00 /usr/openv/netbackup/bin/vnetd -standalone
root     23463     1  0 10:13 ?        00:00:00 /usr/openv/netbackup/bin/bpcd -standalone
root     23493     1  0 10:13 ?        00:00:00 /usr/openv/netbackup/bin/nbftclnt


Shared Symantec Processes
-------------------------
root     23350     1  0 10:12 ?        00:00:00 /opt/VRTSpbx/bin/pbx_exchange

I tried both rebooting the machine and restarting services. No progress yet.
 

[ldias]

Hi guys,

I haven't solved problem yet.

Just help me understanding "NetBackup 7.x Hardware Compatibility List" from May 17, 2013.

On section "Fibre Transport Media Server HBAs"

First it points Red Hat Enterprise Linux 6 on x86-64 Update 1 at "Fibre Transport Media Server - Supported Platforms" section as Supported.

Then at next section "QLogic - Windows / Linux" it points that if I have an HBA of SANblade 24xx series

then I have to use Red Hat 4 or Red Hat 5, is that correct ? I got a bit confused.

[Juannillus]

Hello,

Something similar happened to me in the past. The storage guys told me that the zone was properly configured, but as time demonstrate later on it wasn't correctly configurated. What I did was to review the creatin of the zone myself and I found out that it wasn't correct. You have to do a zone between one client (and one client only) and the FT media HBAs. Could you confirm that this is correct?

Next step that I would do, it would be to increase the verbose level of nbftclnt (originator id 200). Then I would remove the ft client from the master server, and I would add it again, and rescan it.

- nbftconfig -deleteclient <client_name>

- nbftconfig -addclient <client_name>

- nbftconfig -rescanclient <client_name>

If it doesn't work, please provide us  with the nbftclnt log with verbose at 6.

Regards,

Juan

[ldias]

Hello ,

@Juannillus the zoning is correct. It's 1:1 (client and ft media hba).

We are having progress on case opened at Symantec support. They have found the problem.

For some reason nbftclnt services on client discover ft pipes and associates ft media hostname instead of ft media server netbackup name.

According to backline support

FT is enabled on the media server - the media server communicates with EMM and says "I'm active for FT" and EMM registers the name it knows that media server as for FT

Then when the nbftsrvr process runs and queries the local system name on the media server, it passes that name to the nbftclnt process via the Archive Python device

The nbftclnt process queries all SCSI devices looking for SYMANTEC PIPE

And this same inquriy gets the media server name presenting the device (client could see FT devices from multiple media servers)

Client then goes to EMM and says 'I found FT devices for media server hostameyyyyy

And EMM says 'I don't have a media server with that name active for FT.

11/06/2013 12:10:00.757 [DiscoveryTaskThread]  DiscoveryTask Thread 2128 Startup
11/06/2013 12:10:00.772 [AddDevice] \\.\Scsi10:0.0.0
 Inquiry "SYMANTECFATPIPE 0.0     hostamexxxx"
 TargetHBA:LUN:InitiatorHBA = 0:0:0xa0 State = 1 RefCount = 0
11/06/2013 12:10:00.772 [AddDevice] \\.\Scsi10:0.0.1
 Inquiry "SYMANTECFATPIPE 0.1     hostnamexxxx"

Changing hostname solved the problem.

But there's an issue yet. It didn't work on windows client with netbackup 7.5.0.4. Installing netbackup 7.0 on client without any change make it works.

On linux client 7.5.0.4 are running fine.

So the case is still opened to diagnose issue on windows client with 7.5.0.4.

As soon as the problem get completely solved I post it here.

[Juannillus]

I'm glad your initial problem is fixed. If I can be of any help, feel free to ask.

Regards,

Juan

[ldias]

By the way @Juannillus, what about your san clients throughput ?

On preliminary tests for database machine (nearly 100GB of data) I get speed of 106 MB/s. For fileserver, speed gets down a bit.

Since we have a dedicated backup network for some clients we already get throughput of  100 MB/s through LAN.

Please tell us about your experience with san client

[Juannillus]

Hello Idias,

we do not have a dedicated backup network, so the increase in the speed was quite big in some clients. Other clients where trasformed from EC into SAN clients, and although we got an increase in speed in every machine, they weren't as great as the "normal" clients converted into SAN clients.

The speed ratios that we get varies quite a lot from one machine to another. The maximun speed that we got with standard policy was up to 137Mb/s, but the average I would say that is around 100-105 Mb/s. Some of the machines get speeds quite slow, but always better than through LAN.

Regards,

Juan