32bit RHEL 5.9 client backup to 64bit NetBackup Server v7.x...

[sdo]

Hi,

I have an N5220 master/media appliance running v2.5.2 (with NetBackup v7.5.0.5) and it can backup a 32bit RHEL 5.9 with NetBackup Client v6.5 (no patch) just fine.

However, when I bptestbpcd from a 64bit RHEL 6.4 master/media running NetBackup Server v7.6.0.1 - I get:

...
16:30:18.400 [23213] <2> put_string: cannot write data to network:  Connection reset by peer (104)
16:30:18.400 [23213] <2> bpcr_put_vnetd_forward_socket: put_string /tmp/vnet-blah-blah failed: 104
16:30:18.400 [23213] <2> local_bpcr_connect: bpcr_put_vnetd_forward_socket failed: 24
16:30:18.400 [23213] <2> ConnectToBPCD: bpcd_connect_and_verify(xxxxx, xxxxx) failed: 24
<16>bptestbpcd main: Function ConnectToBPCD(xxxxx) failed: 24
16:30:18.400 [23213] <16> bptestbpcd main: Function ConnectToBPCD(xxxxx) failed: 24
<16>bptestbpcd main: socket write failed
...

Is anyone out there successfully backing-up a 32bit RHEL 5.9 NetBackup Client v6.5.x (patched) client to NetBackup v7.6.x.x ?

If so, which v6.5.x patch version works for you?

Many thanks.

[Marianne · Accepted solution]

Add -debug -verbose to the command to see which ports are attempted. Ensure bpcd log folder exists on the client to see if connection attempt is getting to the client.

[revarooo]

Enable logging as suggested and Try from the master:

telnet <client> 13782

Does that work?

[sdo]

TY Mrs M and Mr R...

I have upgraded client to v6.5.6.

But still get same error:

[root@rdcppspnbkp001 ~]# bptestbpcd -verbose -debug -client xxxxx
21:05:44.701 [22423] <2> bptestbpcd: VERBOSE = 0
21:05:44.715 [22423] <2> read_client: dname=CO_0, offline=0, online_at=0 offline_at=0
21:05:44.715 [22423] <2> read_client: dname=.., offline=0, online_at=0 offline_at=0
21:05:44.715 [22423] <2> read_client: dname=., offline=0, online_at=0 offline_at=0
21:05:44.715 [22423] <2> read_client: dname=host_info, offline=0, online_at=0 offline_at=0
21:05:44.715 [22423] <2> read_client: dname=OA_0, offline=0, online_at=0 offline_at=0
21:05:44.720 [22423] <2> db_freeEXDB_INFO: ?
21:05:45.385 [22423] <8> async_connect: [vnet_connect.c:1700] getsockopt SO_ERROR returned 111 0x6f
21:05:45.397 [22423] <8> vnet_vnetd_service_socket: [vnet_vnetd.c:1998] VN_REQUEST_SERVICE_SOCKET 6 0x6
21:05:45.397 [22423] <8> vnet_vnetd_service_socket: [vnet_vnetd.c:2012] service bpcd
21:05:45.447 [22423] <2> logconnections: BPCD CONNECT FROM 172.2.1.1.46824 TO 172.1.2.206.13724 fd = 3
21:05:45.448 [22423] <8> async_connect: [vnet_connect.c:1700] getsockopt SO_ERROR returned 111 0x6f
21:05:45.459 [22423] <8> do_vnetd_service: [vnet_connect.c:2038] connect VNETD CONNECT FROM 172.2.1.1.33990 TO 172.1.2.206.13724 fd = 4
21:05:45.459 [22423] <8> vnet_vnetd_connect_forward_socket_begin: [vnet_vnetd.c:443] VN_REQUEST_CONNECT_FORWARD_SOCKET 10 0xa
21:05:45.499 [22423] <8> vnet_vnetd_connect_forward_socket_begin: [vnet_vnetd.c:460] ipc_string /tmp/vnet-15545424293545708260000000000-ix8YNW
21:05:45.540 [22423] <2> put_string: cannot write data to network:  Connection reset by peer (104)
21:05:45.540 [22423] <2> bpcr_put_vnetd_forward_socket: put_string /tmp/vnet-15545424293545708260000000000-ix8YNW failed: 104
21:05:45.540 [22423] <2> local_bpcr_connect: bpcr_put_vnetd_forward_socket failed: 24
21:05:45.540 [22423] <2> ConnectToBPCD: bpcd_connect_and_verify(xxxxx, xxxxx) failed: 24
<16>bptestbpcd main: Function ConnectToBPCD(xxxxx) failed: 24
21:05:45.540 [22423] <16> bptestbpcd main: Function ConnectToBPCD(xxxxx) failed: 24
<16>bptestbpcd main: socket write failed
21:05:45.557 [22423] <16> bptestbpcd main: socket write failed
<2>bptestbpcd: socket write failed
21:05:45.557 [22423] <2> bptestbpcd: socket write failed
<2>bptestbpcd: EXIT status = 24
21:05:45.557 [22423] <2> bptestbpcd: EXIT status = 24
socket write failed

What I have noticed is that the RHEL 5.9 NBU v6.5.6 client has the same IP on two interfaces (eth0 and eth1) - and doesn't look like a traditional bond to me.

ssh from master to client works ok.

Have tested this from client side, and it works:

xxxxx root :: IX >bpclntcmd -pn
expecting response from server 172.2.1.1
xxxxx xxxxx 172.1.2.206 34623

Add client can telnet back to 13724 and 13782 of the new RHEL 6.4 NBU v7.6.0.1 master.

Unfortunately I don't have telnet or nc on the new master yet.

I'll check the client side logs tomorrow.

[revarooo]

- What I have noticed is that the RHEL 5.9 NBU v6.5.6 client has the same IP on two interfaces (eth0 and eth1) - and doesn't look like a traditional bond to me.

Definitely get that looked into - that could be causing some big networking issues

[sdo]

In the end a typical issue.  Me making assumptions.

Turns out the client was unable to reverse lookup the master/media server; no reverse DNS address, and no hosts file entry.

Marianne was right, the bpcd client logs showed it with a:

<16> bpcd peer_hostname: gethostbyaddr failed to return peer host, herrno = 1

Resolved by adding a hosts file entry.