Cannot log into my HP machine

[Nicky_5]

Hi All,

I had posted a thread on this earlier too and thought the problem was solved.

I have an HP machine running HP-UX 11.0.
the machine is hardly used, but every once in a while i try to login (using telnet / rlogin / ssh) from my laptop (196.143.12.20)
i am unable to login, it does not give me any error (like permission denied or times out etc)
Also when i ping form my laptop i get the following message( note that the loss is 0%)

ping 196.143.12.32

Pinging 196.143.12.32 with 32 bytes of data:

Reply from 196.143.12.32: Source quench received.
Reply from 196.143.12.32: Source quench received.
Reply from 196.143.12.32: Source quench received.

Ping statistics for 196.143.12.32:
Packets: Sent = 3, Received = 3, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 0ms, Maximum = 0ms, Average = 0ms

Also, i tried the ping form another HP machine
( 12.20.194.12) and the ping output was again with 0% loss but this time it was a normal ping output. But again i am unable to login to the machine even from thie other machine.

When i had posted this problem the last time i got a response that maybe the inetd daemon was dying. So i put in a cronjob that runs every 5 minutes and checks to see if inetd is alive. If it is not then it will start it and send out an email to me. However i have received no email form this machine which goes to show that this is not because of inetd dying.

Please advice.

[HGN]

Hi

You can refer to the following 2 thread's

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=91425

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=689778

This should answer your question.

Rgds

HGN

[Nicky_5]

Hi,

Thankyou for your reply, How ever my problem is not the ping results. Pings form my PC to the problem machine works fine when i am able to log into the machine. There is something thats happening on the server itself thatis causing this to happen every few days. If i reboot my machine, everything will be fine. i will be able to login and the pings will also be normal.

Can someone help me out here... Is there something i can do before rebooting the machine? i dont want to reboot the machine yet, because if i do, all traces would vanish and i will not be able to find out what caused this.

[Rick Garland]

Seems as if the bigger problem is not being able to login. The system reports it is up, you can ping the system, but will not let you login. You can have a login prompt but it just cycles around the prompts, never letting you into the system, as any user.

The source quench messages I can ignore for now.

Had a symptom similar to this but cannot find my notes related to this issue. It was related to a bad disk drive, the disk that the root filesystem was on.

[Tim Nelson · Accepted solution]

I would initially guess duplicate IP.

Try some of the below from your other server(s).

Send a broadcast ping, then arp -a. Check mac addresses with the server having problems. See if it is correct or if changes.

Take server one down. Try pings from other server and/or PC. If something responds then you got yourself a dup IP problem.

Change IP on server.

[Tim Nelson]

Once you find the cause post the solution.

Also, I noticed through your profile that you have only responded to 26 of 153 questions with points.

Show your cohearts your appreciation and assign points. You may find the answers come faster.

[Steven E. Protter]

I'd back up the duplicate ip idea.

I'd go to the console or log into the webconsole and change the ip address with the ifconfig command and see if the problem is resolved.

inetd has nothing to do with not receiving email. Email goes out through the network and if there is a problem with the network it queues up on the machine.

If you prove the duplicate ip hypothesis, contact network administration and either get a new ip address or get the usurper off the network.

SEP

[Nicky_5]

Hi,

I do not even get a login prompt. If i do a telnet form dos prompt, i just get a blank screen that goes on forever (no prompts, no timeouts)

If i login using an emulator called secure CRT, again i just get a blank screen as above. And if i try to close the session by closing the window it asks me if i want to disconnect from the server (i usually get this message if i try to close a sucessfully connected session. If try to close a session thats not connected then i dont get any message).

if this was a hard disk problem, why is it that i am able to login once i have rebooted the machine.

can someone suggest something that i can do / check form my end? i have run out of ideas...

[Mel Burslan]

It sounds like there is a workstation sharing your server's IP address which should not be very hard to reboot, especially if it is running the Redmond's fines OS, this is the only way to troubleshoot. When you boot your server, late comer grabs the IP address but it seems like the other user of the same IP is rebooting his/her box soon after yours come alive and you are losing your connection due to the "late comer" principle.

To diagnose this, shut down your server and while it was down, ping the IP address. If the IP resonds to ping requests, you know your problem. Contact your network admins to pinpoint the rogue box on the network and have a talk with the person who gave them permission to use this IP address.

[Nicky_5]

Hi Tim,


Thankyou for your reply, i am sorry about the points, but if you see my last few threads, i have assigned points to every response. Like you another member told me about the points, since then i have been assigning points to all response.

About my problem at hand, i am sorry if this sounds silly, can you tell me how i can send a broadcast ping?

i ran the command arp -a form 2 machines (one in the same zone as the problem machine) and saw the entry:
luna.domain.com (196.143.12.32) at 8:0:9:b8:79:1e [ethernet] stored in bucket 59
bucket: 59 contains: 2 entries

and the command arp -a from another machine in another zone did not have any entry for this machine.


Now what do i do with this info?

Also form Mels response, i shutdown my machine and tried to ping the machine. But i got a request time out which i guess goes to prove that another work station is not using this ip.

Please advice what i should do now. The machine is still down, let me know if there is something that i should do while the machine is down, or if i can bring up the machine?

[Tim Nelson]

broadcast ping.

172.20.10.255 (if class C subnet and mask)

This will ping every server on the 172.20.10 subnet. Depending on the answers you will end up with ARP entries for each.

Interesting that no server answers ping when server1 was taken down. We all were most sure of duplicate IP.

Maybe post some outputs of the following from server1 (bad one):
netstat -in
netstat -rn
/etc/hosts entries that are related to this IP(s).

[Rick Garland]

The dupicate IP you have ruled out.

The system in question was powered cycled? If so, do you have a login prompt now? Can you login?

If the answers are 'YES' this could point to bad root disk or bad GSP. To fix a call was placed to HP.

[Nicky_5]

Hi All,

The server is booting up. Once it comes up i will post the output of the commands you have requested.
In the mean while i was still thingking of what Mel explained.

As He explained, the latecomer machine will grab the ip address, so may be in my case some work station booted, grabbing the IP address of the server, and causing my server to be up but with out an IP(?). Now what happens if the workstation that grabbed the server ip is shutdown. Will my server get its IP back? or will it get it back only after a reboot?

When i did a ping i got a time out, may be the workstation was also down at that time. I dont want to rule out the possibility of duplicate IP yet. Is there any other way to find this out..

[Tim Nelson]

I would assume that the routers, if any, and if on separate subnet, will have to expire their learned arp entries before the first server would come back into the mix.

One way to possibly force it would be to ping from server1 to some other servers on other subnets, you would have to do this from the console.

[Nicky_5]

Hi,

the machine is now up. I am posting the output below:

netstat -in
Name Mtu Network Address Ipkts Opkts
lan0 1500 196.143.12.0 196.143.12.32 3824 1919
lo0 4136 127.0.0.0 127.0.0.1 2355 2355

netstat -rn
Routing tables
Dest/Netmask Gateway Flags Refs Use Interface Pmtu
127.0.0.1 127.0.0.1 UH 0 2355 lo0 4136
196.143.12.32 196.143.12.32 UH 0 1295 lan0 4136
196.143.12.0 196.143.12.32 U 2 0 lan0 1500
127.0.0.0 127.0.0.1 U 0 0 lo0 4136
default 196.143.12.1 UG 0 0 lan0 1500

Please let me know if you want me to post any other output in order to debug this problem

[Mel Burslan]

when you lose your connectivity, as your network admins to do a forced arp on all different subnets and/or vlans from the router level to this IP address and see what MAC address responds and if more than 1 mac address responds to the arp request on two different network segments. Unfortunately arp does not travel through routers and some layer 2/3 switches. So you can not do the arp yourself and expect to get a meaningful result.

if worse comes to worst, you can ask network admins to put a sniffer to the segment of your workstation and you try to telnet to this server. They should see at that point which mac address it is establishing the connection if any.

[Tim Nelson]

Check on Mel's suggestion and while you are waiting on your lan group.

I assume after the reboot all is swell ?
If possible change the IP on server1 to something other than 196.143.12.32.
Depending on your organization you may need to as LAN for a different unused IP. or review what comes back from your broadcast ping (ping 196.143.12.255, assuming class C mask) and choose a number that is not used.
It may be possible that if some other computer is running with 196.143.12.32 then they are having problems also and they keep rebooting or shutting down just like you.

If your organization keeps dns updated you could attempt to "nslookup 196.143.12.32" to see if a name other than your server comes back (do this from some other server or PC). This may lead you to the culprit.

[Sheriff Andy]

I have to agree with some of the other comments above about there being a duplicate IP out there on the network. You need to harp at your DNS administrator to get him to correct the problem.

The nslookup command on the IP like someone else suggested might shed some light as well.

Andy

[Nicky_5]

Hi All,


I have this problem again. The last time when i posted this and rebooted the macine, all was fine. Now i am unable to connect again!!..

I really dont understand what the problem is... Can it be that for some reason the Machine is hanging? How can i diagnose such a situation?

As per Mel's and Tims response, i did the following:

1)
nslookup 196.143.12.32
gave the correct machines DNS name (did not come up with any other name)

2)
ping 196.143.12.255
Pinging 196.143.12.255 with 32 bytes of data:

Reply from 196.143.12.2: bytes=32 time<1ms TTL=255
Reply from 196.143.12.45: bytes=32 time<1ms TTL=64
Reply from 196.143.12.103: bytes=32 time<1ms TTL=64
Reply from 196.143.12.45: bytes=32 time<1ms TTL=64

Ping statistics for 196.143.12.255:
Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
Minimum = 0ms, Maximum = 0ms, Average = 0ms


I dont know what else to do... please advice..

I am trying to get to change to IP address of the machine to see if that solves the problem. My network admin believes that this nothing to do with IP, She says that the machine is probably hanging..

Please advice..

[Rick Garland]

I am voting for the machine hanging.
This sounds like the boot disk/GSP hanging.

Seen this before. System starts OK, after some time the system will still respond to pings but will not allow logins.

Reboot system and is fine again until next occurance...

[Nicky_5]

Also, the forced ARP done on the subnet by my network admin, came back with only 1 MAC (thats the one of my machine).

[Nicky_5]

Hi Rick,

Can you tell me how i can diagnose this? i know i have to reboot the machine. but can you tell me the steps to mke sure that the boot disk/GSP is hanging.

I am an AIX Admin and not very comfortable with the HP commands..

Thankyou..

[Rick Garland]

The issue was diagnosised via HP support. Sorry I do not have the specifics in hand as we speak but access into the GSP was seen to be impacted. Sometimes couldn't get in at all.

[Nicky_5]

Hi,

Thankyou for your reply. But isnt there a command or logfile which shows us the hardware & software errors on the system?

Like in AIX we have errpt that logs all errors in the machine...

Also would you know if this kind or problem (that is the Boot / GSP hangind would be an error that would be logged by any error daemon?

[Tim Nelson]

Not sure if we ever confirmed this.

When this problem occurs you can log into the console with no problem ? I am assuming this as rebooting the system did not seem to be a problem.

If the console login works but the network is intermittant or does not work then the system is NOT hung.

If there is a nic issue then reviewing the nettl00x.log in /var/adm/ may help to detect any LAN disasters, /var/adm/syslog/syslog.log may lead you to places also.

You could possibly have a failing nic or network problem ( if switched network as LAN guys to check for port errors). You can use lanadmin to look as well for frame or link errors.

Another check. could be a auto negotiate issue. As LAN group to check setting on switch if set to auto then force both switch and nic to 100FD. ( auto is bad ).