Crash with MemoryError

[kernelzeroday]

Ran into this issue while testing the script. Worked fine a few times, then halfway through cleaning some logs i got a crash like this, then it continues to crash. IP and hostnames replaced as they are not relevant.

root@generic_host_name:~# python nojail.py -i x.x.x.x -n x.x.x.x
[ ] Cleaning logs for sjkgxd (x.x.x.x - x.x.x.x).
[ ] No entries to remove from /var/run/utmp.
[ ] No entries to remove from /var/log/wtmp.
[ ] No entries to remove from /var/log/btmp.
Traceback (most recent call last):
  File "nojail.py", line 601, in <module>
    clean_lastlog(LINUX_LASTLOG_FILE, args.user, args.ip, args.hostname)
  File "nojail.py", line 331, in clean_lastlog
    clean_file += f.read()
MemoryError

Any ideas? The system was nowhere close to out of memory.

Thanks

[JusticeRage]

Wow, the /var/log/lastlog file must be enormous to cause such an error. Would you be able to give me its approximate size?

Also, as I cannot reproduce this error on my end, could you apply the following patch and verify that:

1. It solves the issue you have
2. The lastlog update still works as intended?

--- a/nojail.py
+++ b/nojail.py
@@ -328,12 +328,15 @@ def clean_lastlog(filename, username, ip, hostname):
                                                              LAST_LOGIN["hostname"])

         # Append the rest of the file and overwrite lastlog:
-        clean_file += f.read()
         tmp_file = get_temp_filename()
         g = None
         try:
             g = open(tmp_file, "wb")
             g.write(clean_file)
+            lastlog_end = f.read(4096)
+            while lastlog_end:
+                g.write(lastlog_end)
+                lastlog_end = f.read(4096)
         finally:
             if g is not None:
                 g.close()