I'm trying to use tail/grep to monitor a log file. The command I cooked up is:

tail -n 50 -f output.log | grep 'type:system' | cut -f 5-

A sample line from the log file is:

1208894862 type:system session:0 severity:4 load started

the columns are tab delimited.

this works ok, except that it's inexplicably lagging badly. If I have 2 terminals open, one simply tailing the file and the other tail|greping the file as specified above, it might take 30 seconds or more for the tail|grep terminal to display the next line while the tail terminal displays it instantly.

EDIT: forgot to add this.
Also, sometimes new text does not appear until even more txt is added to the log file.
Example:
'test 1' is written to the log
tail will show this right away, tail | grep will NOT
'test 2' is written to the log
tail will show this right away, tail | grep will show both 'test1' and 'test2' now



Any hints for me? Thanks in advance

The pipe buffers I/O so it won't be printed until the output buffer is full. This is a classic FAQ. Google for "grep unbuffered" for example.

Thank you for your reply, it's much appreciated.

I think I understand the issue, but I don't see an obvious solution in your reply. Should I not be using a pipe? Is there some grep option I should be using, or maybe awk or sed instead?

Some grep implementations have a --line-buffered option which would help. I don't think there is a standard, one-size-fits-all solution to this. Maybe a Perl script to do the tail and grep in the same process; that should certainly avoid any pipe buffering.