Socket error

[Rithesh]

Hi all,

I have a sql cluster in which one of the machine fails to take back up

The server shows status code 58 "cannot connect the client"

When i checked the host in admin console it fails to connect by showing error code 25 "socket error"

I made the basic troubleshooting & all seems fine.

Master server is NB 6.5.2

client is windows server Ent x64.

Below is the bpcd from client.

00:02:02.722 [46504.39176] <4> bpcd valid_server: hostname comparison succeeded
00:02:02.925 [46504.39176] <2> bpcd main: output socket port number = 1
00:02:05.738 [46504.39176] <2> get_long: (2) premature end of file (byte 1)
00:02:05.738 [46504.39176] <2> get_vnetd_forward_socket: get_string ipc_string failed: 5
00:02:05.738 [46504.39176] <16> bpcd main: get_vnetd_forward_socket failed: 23
00:02:10.379 [43292.46884] <2> bpcd main: offset to GMT 25200
00:02:10.379 [43292.46884] <2> bpcd main: Got socket for input 536
00:02:10.379 [43292.46884] <2> logconnections: BPCD ACCEPT FROM 10.125.1.66.1140 TO 10.125.6.5.13724

Bprd from server

vnet_connect_to_vnetd_extra: vnet_vnetd.c.179: msg: VNETD CONNECT FROM 10.125.1.66.1593 TO 10.125.6.2.13724 fd = 1724
01:54:22.920 [4052.5784] <2> vnet_pop_byte: vnet.c.184: errno: 10054 0x00002746
01:54:22.920 [4052.5784] <2> vnet_pop_byte: vnet.c.186: Function failed: 43 0x0000002b
01:54:22.920 [4052.5784] <2> vnet_pop_string: vnet.c.266: Function failed: 43 0x0000002b
01:54:22.920 [4052.5784] <2> vnet_pop_signed: vnet.c.310: Function failed: 43 0x0000002b
01:54:22.920 [4052.5784] <2> version_connect: vnet_vnetd.c.1812: Function failed: 43 0x0000002b
01:54:22.920 [4052.5784] <2> vnet_connect_to_vnetd_extra: vnet_vnetd.c.183: status: 43 0x0000002b
01:54:22.920 [4052.5784] <2> vnet_connect_to_vnetd: vnet_vnetd.c.149: Function failed: 43 0x0000002b
01:54:22.920 [4052.5784] <2> nb_vnetd_connect: comm.c.1887: vnet_connect_to_vnetd failed: 43
01:54:22.920 [4052.5784] <2> bpcr_vnetd_connect_forward_socket_begin: nb_vnetd_connect(10.125.6.2) failed: 25
01:54:22.920 [4052.5784] <2> local_bpcr_connect: bpcr_vnetd_connect_forward_socket_begin failed: 25

01:54:22.920 [4052.5784] <2> ConnectToBPCD: bpcd_connect_and_verify(DB02, DB02) failed: 25

Please suggest a valid solution.

Thanks

Rith

[Rithesh]

One thing i want to mention is that once i reboot these cluster servers, then backup will work , but after some days this again occur.

[Srikanth_Gubbal]

Hi,

for cluster you need to back it up by adding the cluster name as client. in this case which ever server is active that server will be backed up.

apart from this if you want to back up both the servers then you need to back up individually by excluding the shared drives of that servers. 

[Rithesh]

Thanks srikanth for the input.

But this machine is in cluster with 2 cluster groups & with two PBX resource group.

I heard that this group get installed default, but two resource is required?

One thing i observed that 2nd pbx resource group keeps on fail which inturn reflect my backup with Status code 24 Socket error.

Please advice asap.

[Criss_R]

Make sure that you have the netbackup client installed on all the physical boxes in the cluster and that all the settings are correct. Then I would backup the cluster using the virtual names in one policy. In another policy, I would backup the physical boxes local drives, not the shared drives. 

[Rithesh]

Hi criss,

As you told i have exactly the same configuration.

But still we face the same issue...intermeittently we can able to connect the client through admin console..but after a few minutes it display with SC 24/25.

Can anyone explain what  is the role of PBX_resource group in cluster which will be created during NB installation.

I am facing the issue when one of this PBX resource goes failover, even though after failover the client fail to connect or shows socket write error.

Thanks

rith

[Rithesh]

Hi all,

Can anyone provide a consistent solution for the above issue....!!!!!!!!!

[J_H_Is_gone]

the pbx is something that was added if your backups are cluster aware

if you do not have or want a cluster aware set up you can remove it.

see tech doc

http://seer.entsupport.symantec.com/docs/304228.htm

we just remove remove the pbx if it gets in stalled (you can also just disable it)