0

We have a POS application that at 4 specific locations will sometimes lock up for 30-120 seconds rather frequently (every several minutes or so). The point in the UI at which it is not responding varies widely. Each location has only the 1 machine running the app and MySQL and whatever else, yet when that machine is brought to our office, the problem does not present itself at all. The only that changed when it was brought here was what network it was connected to.

I'm at a loss for how to troubleshoot this predicament. Any suggestions are welcome, including profiling apps that can tell me what a process was doing when or waiting on when it was not responding.

(And hopefully this is the right site to ask this on since I suspect something network related.)

Edit1 - Additional information for questions asked: OS: Not linux. Windows XP and a cheap router is about how the setups are at these locations.

Cables: I'll have somebody check that.

Connections: It does connect for credit-card processing. One location does not, but in asking about that location I was also told that the presence of this issue is unconfirmed.

Lock-up recovery: The app will resume responding.

Changes to application: It is unconfirmed for now that dropping back to an older version of our application alleviates or eliminates occurrence of the issue. Other locations have the same version without issue.

Edit2: Might not be a network issue after all. I've given everybody some points for trying to help.

Joel B Fant
  • 101
  • 3

2 Answers2

1

a 30-120 sec lock up of a system. Do you need to recover it? IF it recovers by itself after the given time then best check the network connection, might just be faulty cables simple way to test is out is to run a ping -f and have a look at the numbers. If it changes too drastically then chances are its the cable

CChock
  • 156
  • 3
1

I second the network cable test if the network cables are in a situation where they are being moved around like inside of one of those floor pancakes.

Another thing to look at is DNS. If your app relies on DNS for some sort of network resolution and, assuming you don't control DNS, you could augment DNS with host table entries.

Another thing to look at is nailing up your NIC card. If you auto-negotiate, try setting the speed to 100MB-FULL and see if that resolves. If this is linux, try examining dmesg output to see if you see NIC card disconnects/reconnect patterns.

There are other things to look at, but these are the two I would start looking at.

Other things, if running, are /etc/nsswitch.conf settings. Are you running local LDAP? (again, assuming this is Linux variant).

Kilo
  • 1,554
  • 13
  • 21