Originally posted by: SystemAdmin
In the past I've seen this with poor wiring. Sometimes slowing the baud rate down can fix this, sometimes you need better cables or the grounding has been done wrong.
The canned IBM response to this appears to be:
Most PMRs attribute the error message as the result of flow control
problem.
.
The error TTY_TTYHOG indicates that the sending device is ignoring flow
control. This error occurs after the hardware FIFO has been accessed and
has written to the software buffer.
.
In other words, the tty hog error is logged because the buffer where
incoming characters are placed is overflowing. When the buffer is about
3/4's full, the device driver notifies the hardware , to send the XOFF
char to the sender to stop send data (or drop the hardware control line)
. If the sender continues to send data, and the buffer is not emptied,
the device driver flushes the buffer and logs the error message in the
error log.
.
For example, say in a telnet session we enter more than 512 characters
and hit enter. We could fill up our recv buffer on the adapter and the
device driver will send su the XOFF command. However, if we ignore this
and continue to send the block data, we will log this error in errpt.
.
The end point is this is a problem on a client system. It is a temporary
error and there isnt' anything that we will be able to do about if from
a AIX point of view
.
Suggestions
-
check cabling
-
enable flow control on client
-
update bos.rte.tty to last version
#AIX-Forum