cancel
Showing results for 
Search instead for 
Did you mean: 

SAP Backup Receiving Errors (Status 41)

Courtenay_Jones
Level 3
Env:  1X NBU Master/4x Media Servers is 6.0 MP4 - W2K3 -SP1 -LTO3
Client - HPUX - SAP/Oracle Database Agents - 11.23 - PA-RISC
 
When I attempt a SAP backup on this server, it will continually time out with a status 41. The strange thing though is that the cold f/s backups complete perfectly.
 
I checked all of the connectivity with bpclntcmd, and the responses gave back correct from the master.
 
I turned up the logging and in the backint logs, I am seeing the following errors:
 
16:35:45.575 [6295] <8> bsa_checkFeatureId: bprd_get_features() failed, retrying
16:35:45.575 [6295] <2> vnet_cached_gethostbyname: vnet_hosts.c.287: Function failed: 5 0x00000005
16:35:45.575 [6295] <2> vnet_init_connect_rec: vnet_vnetd.c.3623: vnet_cached_gethostbyname failed: 5 0x00000005
16:35:45.576 [6295] <2> vnet_connect_to_service_or_vnetd: vnet_vnetd.c.3499: vnet_init_connect_rec failed: 5 0x00000
005
16:35:45.576 [6295] <2> nb_connect_to_vnetd_or_legacy: comm.c.2033: vnet_connect_to_vnetd_or_service failed: 5
16:35:45.576 [6295] <2> bprd_connect: Cannot connect to server sbs09932bu001.usncraleigh9901.int.wolseNBBSA_CLIENT_H
OST=wrlspd10.usncraleigh9901.int.wolseley.NBBSA_POLICY=p_sap automatically: 11
16:35:45.576 [6295] <2> bprd_connect: Cannot connect to server sbs09932bu001.usncraleigh9901.int.wolseNBBSA_CLIENT_H
OST=wrlspd10.usncraleigh9901.int.wolseley.NBBSA_POLICY=p_sap
16:35:45.576 [6295] <2> bprd_connect:    errno = 0 - Error 0
16:35:45.576 [6295] <2> bprd_get_features: bprd_connect(sbs09932bu001.usncraleigh9901.int.wolseNBBSA_CLIENT_HOST=wrl
spd10.usncraleigh9901.int.wolseley.NBBSA_POLICY=p_sap) failed, cannot connect on socket (25)
16:35:45.576 [6295] <16> bsa_checkFeatureId: unable to check if feature <40> is licensed
16:35:45.576 [6295] <16> bsa_checkFeatureId: Server exit status = 25: cannot connect on socket
16:35:45.576 [6295] <16> VxBSAInit: ERR - The license for the requested feature is not available.
16:35:45.576 [6295] <4> VxBSAGetEnv: INF - entering GetEnv - NBBSA_DEBUGFD
16:35:45.576 [6295] <4> VxBSAGetEnv: INF - returning -
~
 
I am not sure what I am missing here - any suggestions would be appreciated.
15 REPLIES 15

sdo
Moderator
Moderator
Partner    VIP    Certified
Is there a media server involved?  If so, O/S and NBU version?  If not, I assume the master is also acting as a media.
Did you try bpclntcmd from the client to the master?  IP and name?  Twice?  (twice to pick up possible duplicate reverse DNS).
If that's okay, have you installed the correct licences on the master?  Are the licences correct for the client versions?
Each platform/agent/version requires at least one licence - and some platform/agent/version combinations do check the licence count, although most don't.  You can't run v6.x clients against v5.x licences etc etc.

Courtenay_Jones
Level 3
Yes, there are 4 media servers involved. They are all NBU 6.0 MP4 (WIN2K3-SP1).
 
Yes, I tried bpclntcmd from the client to the master.
 
root@wrlspd10:/usr/openv/netbackup/bin> ./bpclntcmd -pn
expecting response from server sbs09932bu001.usncraleigh9901.int.wolseley.com
wrlspd10.usncraleigh9901.int.wolseley.com wrlspd10.usncraleigh9901.int.wolseley.com 10.165.201.30 51267
root@wrlspd10:/usr/openv/netbackup/bin> ./bpclntcmd -pn
expecting response from server sbs09932bu001.usncraleigh9901.int.wolseley.com
wrlspd10.usncraleigh9901.int.wolseley.com wrlspd10.usncraleigh9901.int.wolseley.com 10.165.201.30 51269
root@wrlspd10:/usr/openv/netbackup/bin> ./bpclntcmd -hn sbs09932bu001.usncraleigh9901.int.wolseley.com
host sbs09932bu001.usncraleigh9901.int.wolseley.com: sbs09932bu001.usncraleigh9901.int.wolseley.com at 10.154.106.192 (0xa9a6ac0)
checkhname: aliases:
root@wrlspd10:/usr/openv/netbackup/bin> ./bpclntcmd -hn sbs09932bu001
host sbs09932bu001: sbs09932bu001.usncraleigh9901.int.wolseley.com at 10.154.106.192 (0xa9a6ac0)
checkhname: aliases:
root@wrlspd10:/usr/openv/netbackup/bin> ./bpclntcmd -ip 10.154.106.192
checkhaddr: host   : sbs09932bu001: sbs09932bu001.usncraleigh9901.int.wolseley.com at 10.154.106.192 (0xa9a6ac0)
checkhaddr: aliases:
root@wrlspd10:/usr/openv/netbackup/bin>
Yes, I double checked the licenses on the master - tthey are correct for their version/platform.
We have all 6.x licenses.

sdo
Moderator
Moderator
Partner    VIP    Certified
All servers and clients all talk to each other and they al perform reverse lookup, so the thing I always do, even before attempting a backup of a new client (because the DNS team at my site always forget - so it seems), is to:
 
from client:
nslookup master-name
nslookup master-name
nslookup media-name
nslookup media-name
nslookup client-name
nslookup client-name
 
nslookup master-ip
nslookup master-ip
nslookup media-ip
nslookup media-ip
nslookup client-ip
nslookup client-ip
 
...and then I run all twelve commands again on the media, and then all twelve again on the master.
 
Then after this, I run the equivalent for:
bpclntcmd -hn master-name
bpclntcmd -hn master-name
bpclntcmd -hn media-name
bpclntcmd -hn media-name
bpclntcmd -hn client-name
bpclntcmd -hn client-name
 
and the same size ahain but with -ip <machine-ip> i.e. each IP addy twice (I'm getting bored with typing - but ended up typoing more anyway - sheez!)
 
and then all twelve commands again from media and then from client...
 
 
As you can probably guess I have a little short script to do it for me...
 
It's the only way I know to confirm:
1) No duplicate reverse DNS.
2) No duplicate DNS.
3) No confusion between DNS and hosts files
4) No funnies with WINS
 
 
If things don't work after this try from the master:
 
$ bpgetconfig -M media-name SERVER
$ bpgetconfig -M client-name SERVER
 
and check that for both that the first entry of "SERVER = <blah>" is the master server name.
 
 
 
The try:
$ bpgetconfig -M client-name SERVER
...and:
$ bpgetconfig -M client-name MEDIA_SERVER
 
and check that your media server is listed in one of or both lists.
 
HTH,
Dave.

sdo
Moderator
Moderator
Partner    VIP    Certified
The manuals for the database options mostly recommend a default CLIENT_READ_TIMEOUT of 15 minutes (i.e. 900 seconds) - but the manuals mostly omit to tell you where this setting is important...
 
...and it's not on the client, it's on the media (or master/media if your master is also a media).  So, try checking the "media server client read timeout" with:
$ bpgetconfig -M media-name CLIENT_READ_TIMEOUT
 
...if it's less then 15 minutes (i.e. 900 seconds) then try 15 mins first, then try increasing this, by 300 second increments (5 minutes) until the client works - IMHO if you get to an over 30 minutes and it still doesn't work, then something else is wrong.
 
We do this because some database options take a few minutes (a short while) to quiesce and set themselves up before sending data across.

David_Zhang
Level 4
Partner Certified
Hi,
 
    I strongly suggest that you add one parameter to bp.conf of client as "REQUIRED_INTERFACE = hostname" and "CLIENT_NAME =  hostname".
   And try again.
 
David

Courtenay_Jones
Level 3
David,
 
I can go through and post the results if you like, but as a summary, everything came back clean.
I tried this with using and dns, and just using host files, and the result was identical.
And like I said in the prior post, the cold filesystems backups work perfectly.
 
 

Courtenay_Jones
Level 3
I turned the client timeout option up to 30 minutes, with it still erroring out on a status 41.

Courtenay_Jones
Level 3

root@wrlspd10:/usr/openv/netbackup> cat bp.conf
SERVER = sbs09932bu001.usncraleigh9901.int.wolseley.com
CLIENT_NAME = wrlspd10.usncraleigh9901.int.wolseley.com
REQUIRED_INTERFACE = wrlspd10.usncraleigh9901.int.wolseley.com
MEDIA_SERVER = sbs09932bu002.usncraleigh9901.int.wolseley.com
MEDIA_SERVER = sbs09932bu003.usncraleigh9901.int.wolseley.com
MEDIA_SERVER = sbs09932bu004.usncraleigh9901.int.wolseley.com
MEDIA_SERVER = sbs09932bu005.usncraleigh9901.int.wolseley.com
VERBOSE = 5
ENABLE_PERFORMANCE_DATA_COLLECTION = NO

J_H_Is_gone
Level 6
This might not be it - but I just spent 3 months fighting 41's
 
turns out the write cache was off on the san drives, which cause the data to just be too slow for the backups to work correctly.
 
so you can check that, might not be it, but it is something to check.

Courtenay_Jones
Level 3
I checked and the write cache is turned on,  but thank you for the suggestion.

js88699
Level 5
Partner
Have you tried to run the scripts on the SAP side and see what happens?
 

sdo
Moderator
Moderator
Partner    VIP    Certified
Have you seen the "In depth troubleshooting of status 41 for version v6.0"?
 
 
Also, have you tried FTPing a really big file from client to media and media to client?
 
 
Also, if you have Symantec Support arrangements you could open a case and they'll get you to run an ApareNet sequencer tool to check out the NIC and drivers and O/S interaction.
 
 
Personally I have seen status 41's caused just by having out of date NIC drivers.

sdo
Moderator
Moderator
Partner    VIP    Certified
Looking at it again, I see this in your logs:
 
16:35:45.576 [6295] <16> bsa_checkFeatureId: unable to check if feature <40> is licensed
16:35:45.576 [6295] <16> bsa_checkFeatureId: Server exit status = 25: cannot connect on socket
16:35:45.576 [6295] <16> VxBSAInit: ERR - The license for the requested feature is not available.
 
 
Where did you install the licences, client, media or master?  I believe the licences have to be installed on the master.  You say the file system backups run fine, so it really looks like the client has an issue with licencing.  Maybe it's actually a new bug?  If you're sure that you've got your licencing right then I think you're going to need to log a case with Symantec Support - as this is a user based forum, and not an official product support forum.

Message Edited by David Rawson on 08-03-200709:39 PM

sdo
Moderator
Moderator
Partner    VIP    Certified
Just noticed this fixed bug in the v6.0 MP5 release notes for HPUX server:
 
 
Etrack Incident = ET574945

Description:
  Snapshot client SAP backups would fail because no valid license key be
  found on the server.
 
 
The fix is also listed in the v6.0 MP5 Unix (all flavours) client pack too.
Could this be related to your problem?

Message Edited by David Rawson on 08-04-200705:03 PM

Stumpr2
Level 6
does SAP have its own clientname instance?