Topic: Intermitent Server Access Problems

[ask_jeeeves]

Hi,

I wondered if somebody could help me to identify what i have done wrong.
The last 3-4 days, i have had trouble being able to access things locally on
my SME Server. It seems that the connection locally to the server keeps dropping, for
3 -4 minutes then its fine for say 2 minutes then down again.
This includes being able to view ibays, webpages, ssh, and making imap connections locally.
But when i cant connect to server manager i can go to external ip which points to
my primary ibay, plus i can go to www.domain.xxx  and that all works absolutly fine.

I use Thunderbird by Mozilla and have trouble with an
error message popping up. I Increased the ConcurrencyLimitPerIP in this
HowTo: http://wiki.contribs.org/Email#Email_Clients    which was around the time i noticed this
problem starting. I have changed this back to what it was, but no change.

I am running SME 7.2 in server only mode.

Thanks very much.

Rob Miller

[ask_jeeeves]

Does anyone have any ideas???
I am finding it hard logging in to check anything as i keep getting a 404 connection error
going to server-manager and i cant ssh but i can view the webpages in the ibays
using external doman names and i can ftp externally

Any ideas greatly appreciated

Rob Miller

[christian]

It's not clear to me what your config looks like or what path your requests are taking to get to your server.

It could be connectivity to your server on say the LAN side but I don't know if your requests to your public domain address is taking the same path or a different path to your server. If it is connectivity then start with the easy stuff like changing cables, L2 switch ports, and lastly internal NIC.

But first I would check for rogue processes using htop via a shell window to see if anything is dominating your CPU or memory.

More info will help people help you.

Christian

[ask_jeeeves]

Hi Christian,

I have a small network of 5 computers and SME acting a web & email server.
I have a problem accessing SME locally. I cannot accesses server-manager
through local ip xxx.xxx.xxx.xxx/server-manager, nor can i access ibays locally
or check mail through a client, or make an ssh connection, but when this is happening i can access domains
hosted on the server through http and can connect to webmail from the internet.

Its really weird. And its not anything to do with the ConcurrencyLimitPerIP as i posted before.

I did try the htop command before but nothing comes up out of the ordinary before the ssh connection drops.

When i cant get local access to the server, i can ping it locally.

Can i give any more info?

Thanks

Rob Miller

[christian]

Can you access the server via http using its local DNS name vs its external DNS name? Or even its internal IP address. eg:
http://192.168.1.1. My your 2nd note, I'm guessing it works well only if you access it from external.

What does the LAN topology look like? Is the server acting as your internet gateway too? Are you using two NICs? If you had a bad link to the local ethernet side but a good link on the WAN side it could explain this but it depends on your topology.

What is doing DNS lookup, DHCP, etc.

If you have not done any updates when this started and rebooting doesn't solve the issue then my #1 guess is it is HW or link related.

5 PCs shouldn't drive much of a load so even a light processor should be ok. I run a duron 1GHz with 512MB and it supports quite a lot of devices in my LAN.

Christian

[NickCritten]

Have you tried accessing the server from a different PC on the LAN?

What happens when you do a continual ping from your client to the server?
e.g. (windows) :

Code: [Select]

ping -t 192.168.1.10
What happens when you continual ping from the server to the router? (using keyboard & monitor directly)
e.g.

Code: [Select]

mtr 192.168.1.1 (Let it run for a minute & see if you get any Loss)

I would suspect an issue with your windows client / NIC / Switchport

Nick

[ask_jeeeves]

Hi Christian,

Yes when it's all running OK, i can access the server using a local ip address, and
it's DNS name. But when the problem happens, i can't.
I am running in server only mode, and only have 1 network card. There is no DHCP on
the network as all computers etc have a static ip address, so DHCP is switch off on the router
and SME.
SME is up to date.

Rob

[ask_jeeeves]

Hi Nick,

When pinging the server from a workstation locally, i get a good reply
and also pinging the router from SME without loss, but when the problem starts,
it just stops.

Rob

[NickCritten]

Hi Nick,

When pinging the server from a workstation locally, i get a good reply
and also pinging the router from SME without loss, but when the problem starts,
it just stops.

Rob

So when you are having problems, your server can't ping the router? AND at the same time your workstation can't ping the Server?

This is a bit weird as if your server can't ping the router then you shouldn't be able to access it from outside (as the traffic comes in through the router).

Are you sure that this is the case?

[christian]

Sorry to disappear on you. I had a family emergency.

When you say you are accessing your server using your external IP address, are you doing this from another physical entry point to the internet (ie. someone else's house)? Or are you doing this internal to your LAN?

Did you do any updates or add any rpms or contribs to your server when this started?

If you can access without problem from external then you should try and do your diagnostics from the external address. So perhaps SSH in from external and run htop to see what is going on during the "down" time.

You still haven't told us what is your physical layout for your network. I'm guessing you are using a low end Router/NAT (like Linksys or Dlink). It is likely 8 ports or you have two switches in the network to support 6 NICs. You have the router port forwarding to your SME possibly settting DMZ. All other PCs are connected to the router directly. Do you have any switches in between the PCs and the router?

My best guess based on the above assumptions is that you have a problem with the routing layer or the Layer 2 switch in the router. Have you tried replacing it? It is also possible you have something wrong at the Routing protocol layer on the LAN side (doubt it though). If it was at the protocol layer then it may be some piece of equipment on your LAN (possibly another PC, Router, Media Player, etc) that is saturating your LAN. Have you checked the activity lights on the Router and Switches to see what happens when you can't access your server? Have you tried disconnecting all other equipment on your LAN from the router except one PC and the server?

To me it seems unlikely to be your SME server based on the behaviour so just go through a process of elimination as per above and see if you can narrow it down.

First I would look at HTOP via external address (since you say external works). I would then disconnect everything but the Router, SME server, and one PC. no other equipment, not even another L2 switch.  If it is still happening then try substituting the Router, the access PC, the cables, the NICs.

Let's see where we get to with this.

Christian

[ask_jeeeves]

Hi Christian,

I found the problem. I have a LinkSyS PAP2 Vonage Box for VoIP, and for some reason, it decided to use the same IP address as the server.
I am not sure how this happened, as the router and SME do not provide DHCP, as everything here is assigned manually.
But anyway, with the LinkSyS box unplugged, everything is back to normal.

Thank You

[christian]

Good stuff. That would do it. But I wonder why it was intermitant?

[ask_jeeeves]

I don't know. Strange to me. It's as if it tried to connect with the same IP Address as the SME box, then fail and give up for a few minutes, then try again, each time knocking SME Offline.

Rob