Long ago in a galaxy far away I ran into an explanation about lines in logs that are nothing but repeated ^@^@^@ characters. Sadly I never copied that explanation into my notes. Searching for related discussions these days seems futile.

My guess is the characters are escaped characters that render as ^@^@^@ in text viewers.

* What causes these obscure log entries?
* Is there a way to search logs for these obscure characters?

Often these log entries are discovered when using grep with a response the "binary file matches."

Thanks.

Messages were logged, which extended the length of the logfile. But the system crashed before the filesystem cache was written to the physical drive. Consequently, after a reboot, the logfile contains 0-bytes (rendered as ^@) where the messages would have been. This is the scenario in which I encountered the problem.

1 members found this post helpful.

Pressing control key subtracts 0x40 from the ASCII code, code of "@" is 0x40, so ^@ is 0x00 - these are zero (0x00) characters (which usually indicates some sort of file corruption).

That makes some sense. I have seen these characters in syslog, but today I found the characters in maillog. The time stamps in the maillog indicate no system crash. Perhaps I had restarted sendmail at that point but one would think stopping the service would be graceful.

That info led me to change my search criteria. These weird log entries are null characters?

Yeah, I can understand this happening once in a blue moon with syslog, but maillog? Kind of weird.

So how to find the exact line in the affected log with these corrupt strings? Looks like the following might succeed:

grep -Pna '\x00'