11-09-2011 11:32 AM
Master: Solaris 10
Client: Windows Server 2008 R2
NetBackup: 7.1.0.1
-
Some of my jobs are failing with error 58 and when I try to ping, it fails. However, sometimes i do get ping and I've noticed whenever I do, the jobs backsup successfully. I mean, how would I do the troubleshooting when I'm not even getting the ping? What i noticed, all fails when ping fails. However, ping does work sometimes, and I mentioned this to the network team, and they said, it's all good from there end so now you need to get the owner of that master server to do the work? What does this mean? I have full access to this Solaris10 master server, can someone please guide me? and yeah, I assure you when the ping works, i get all bpclntcmd and whatever correct and working fine. it's just when I don't get the ping, i get 58. and I have also noticed, it tries 3 times, and sometimes its fails on 58 but on retry it backsup successfully.
-
The ping from Master server is on/off, very shaky. Client ping working 24/7.
-
Help?
11-09-2011 11:55 AM
What is used on master server for host name lookup - hosts file or DNS?
(grep hosts /etc/nsswitch.conf)
Master and client on same or different subnet?
When ping does not work, can you ping the IP address of the defaultrouter?
What does the routing table look like when ping is not working?
(netstat -nr)
11-09-2011 12:04 PM
My guess is that you have multiple clients with the same IP.
Perform a traceroute when the issue is occurring.
11-09-2011 02:57 PM
if DNS is that flaky then use the hosts files - all master and media ip addresses on every client and every client in the master and media servers host files.
If you really want to test your network use the NetBAckup network and domain analyser:
http://www.symantec.com/docs/TECH125454
Hope this helps
11-09-2011 03:43 PM
Nsswitch on master -> hosts: files dns
Yes, i think they're on different subnet, but am not sure. I don't know the subnet mask so, but i can type the ip address here:
Master: abc.xyz.100.50 Client: abc.xyz.101.120
I don't have the netstat -rn output as of now, but I have seen netstat -a output when It was not working, and I saw it was talaking to media and as well as master on there.
-rn switch will give me the default router right? Do i need that? I would have to ask for it, since i can't get indo a windows box.
We have the DNS thing going on. My take somehow points toward firewall, do you think its possible for the firewall to work and sometimes do not?
11-09-2011 03:56 PM
This is done at master server, and a similar output was coming from all media servers.
Although it gives warning of multiple interfaces but it comes with every x-y-z client-bac, the one which don't have issues.
# traceroute client-bac
traceroute: Warning: Multiple interfaces found; using abc.xyz.100.50 @ e1000g3
traceroute to client-bac (xyz.abc.101.120), 30 hops max, 40 byte packets
1 * * *
2 * * *
3 * * *
4 * * *
5 * * *
6 * * *
7 * * *
8 * * *
9 * * *
10 * * *
11 * * *
12 * * *
13 * * *
14 * * *
15 * * *
16 * * *
17 * * *
18 * * *
19 * * *
20 * * *
21 * * *
22 * * *
23 * * *
24 * * *
25 * * *
26 * * *
27 * * *
28 * * *
29 * * *
30 * * *
11-09-2011 04:01 PM
Where shall I add the host entries? Can you be a little specific? The /etc/hosts of Master & Media, anywhere else?
When i'm running the nettestconn, it is not giving me any output, just a line with 0s. Same with the clients which don't have any issues..
As far as ndna is concerned, well i think i have it installed already, let me know what do now?
11-09-2011 04:02 PM
Just to inform you one more thing.. The host file of the windows machine doesn't have media servers entry there, but I do see media server entries in the registry.
Host files are not importnat when we have DNS, right.
11-09-2011 04:08 PM
I see a host entry in the /etc/hosts of master but it is commented out. I think this is due to the DNS thing. A lot of clients have been commented out, what i conclude from this are the ones which are running 7.1 actually..
11-09-2011 04:19 PM
Every now and again its useful to capture your environment using the nbsu tool. It can gather lots of diagnostic information and you can issue diagnostics on a specific error.
Status code information gathered by the NetBackup support utility (nbsu)
e.g. # nbsu -nbu_e 58
Or just run it your master and go through the files. Can reveal a lot about your environment to yourself.
See my signature for the link to nbsu and also keep the documentation guides close by and refer to them regularly. A LOT of information but familiarise yourself with them relevant to your environment.
11-09-2011 04:27 PM
Where do you want me to run this? On the Client, right?
Well, client doesn't have any problem talking to master. ping, bpclntcmd all works fine. It's the Solaris Master server, which is giving me the problem..
11-09-2011 04:33 PM
On the master.
11-09-2011 04:35 PM
Where would I see the output? in the output folder? It's Solaris10. please confirm me, it wouldn't flood all kind os logs everywhere.. Running it right now.. Just as like you have mentioned.. Is it possible to run this targetting a specific host, like in my case?
11-09-2011 04:40 PM
Ok, i just ran the command. and all i see in the newly created file is some retarded output? What do I make of that file now?
11-09-2011 04:42 PM
Sorry, i thought it was a file. wait, untarring it now..
11-09-2011 04:46 PM
i succesfully gunzip the file, but unpacking tar is giving me this error: tar: /dev/rmt/0: No such file or directory
11-10-2011 12:36 AM
11-10-2011 02:06 PM
Can you run netstat -rn?. Do you set the default router?
11-10-2011 04:04 PM
I'll run this command, can you tell me what we're looking at? Thanks..
11-10-2011 04:05 PM
Which command? Gunzip one? Well for tar am just running tar -xvf <file_name.tar>