Python / read a specific part in a log file

You can use str.split

Demo

>>> content = '''**** start logging ****
... line 1
... line 2
... line 3
... **** end logging ****
... **** start logging ****
... line 4
... line 5
... line 6
... **** end logging ****'''
>>> ''.join(content.split('**** start logging ****')).replace('**** end logging ****','')
'\nline 1\nline 2\nline 3\n\n\nline 4\nline 5\nline 6\n'

You'll need to parse the entire file.

with open(logfile) as f:
    for line in f:
        if line.startswith('**** start logging'):
            l = []
        else:
            l.append(line)
print('\n'.join(l[:-1]))

The last entry in l will be the line end logging, so I explicitly remove it.

Alternatively, you could use file.seek(n, -2), where n is a guess for how many characters each logging record is, but then you would still have to perform some iteration to get the entire record, and check that there is only one instance of the phrase 'start logging' in it, which means you'll execute the code above anyway...

You speak of read and seek so I suppose you do not want to keep a full bunch of lines in memory. You can simply store the position of last logging start and rewind there :

with open(logfile) as fd:
    last_log_start = None
    while True:
        line = fd.readline()
        if line is None: break
        if line.startswith('**** start logging'):
            last_log_start = fd.tell()
    if last_log_start is not None:
        fd.seek(last_log_start)
        while True:
            line = fd.readline()
            if line is None: break
            if line.startswith('**** end logging'): break
            sys.stdout.write(line)