Learn F5 Technologies, Get Answers & Share Community Solutions Join DevCentral

Filter by:
  • Solution
  • Technology

WOM nom nom nom nom

Yes the manic look is something that seems to come with the WOM territory.

WE have just set up WOM on a couple of boxes across a L2 link, everything seemed fine when we had a consultant on site, however when he left we started getting issues.

It does not always optimize traffic, it will ramp up to about 30% utilization of a network connection the drop off again. This is just using the basic FTP profiles. We have run diagnostics on the config and there are no unexpected errors. The only time we get an error, is when we run the network diagnostics when we get a red FAIL error, but no messages pertaining as to what this is.

From what I can see the traffic is not terminating on the server side properly, it passes through the correct VS from the ftp-client it hits the endpoint on the server end, but instead of being terminated and passed through the VS ftp-server, it just goes directly to the physical server, this give the WOM lite optimization speeds but not the full optimization speed we would expect.

These issues only appeared when we turned off De-dupe and tried to create CIFS profiles. We have rolled back to a working configuration which is no longer working, same issues as above, we have also rebooted myriad times.

We are running BigIP version 10.1 WITHOUT the HF1 applied.

Any suggestions or ideas are welcome...well constructive suggestions, as i have already suggested a cramming of the boxes.
Rate this Question

Answers to this Question

Hi: We are trying the same scheme but with version 10.2, as we see everything is going through the iSession tunnel, but the optimization is... invisible, the same time for ftp transfer with or without tunnel, very weird, we can see speeds about 150mbps at the first time (this is link throughput) and ftp client speed is the same, the next transfer at the link the speed goes to 10mbps and the ftp client shows 150mbps, maybe this is the expected behavior but the time of the transfer is the same. Maybe we can talk about our schemes