Hashcat bug
#1
Hi Atom, I have possibly found a bug. I was using Hashcat in a simple dictionary attack with a list of a bit more than 2 million hashes, the Vbulletin external salts and the md5(md5($pass).$salt) algorithm. I was also using the --remove hash function. While the attack was going on (several hours), finding lots of passwords, the temporary hash file got bigger and bigger up to 260 Gb. I have noticed it because my hard drive was almost empty and now, it was almost full.
#2
hi mastercracker, thanks for report. i can confirm this is a bug. as a workaround, do not use -e and --remove in combination :-)
#3
Now that I know this, what I do as a workaround is that after a couple of hours, I manually stop Hashcat, take note of the -s value, delete the hash temp file and then continue the cracking at the -s value.
#4
On a related not to cracking VB hashes, when I specify the salt file which contains 857375 salts, oclHashcat come back as says:

ERROR: too much salts

??
#5
Oh snap, so much salts!
#6
(01-19-2011, 02:08 PM)blandyuk Wrote: On a related not to cracking VB hashes, when I specify the salt file which contains 857375 salts, oclHashcat come back as says:

ERROR: too much salts

??
You have to split the salt file at least in two. And do the attack with each salt file.

#7
Tnx mastercracker, simple enough solution Smile Another issue I have is I only want to use my wordlist files on their own, no right-dict. I need oclHashcat to do this based on the amount of salts but currently I have to use hashcat instead which is slow because of CPU speed Sad
#8
(01-21-2011, 05:51 PM)blandyuk Wrote: Tnx mastercracker, simple enough solution Smile Another issue I have is I only want to use my wordlist files on their own, no right-dict. I need oclHashcat to do this based on the amount of salts but currently I have to use hashcat instead which is slow because of CPU speed Sad
It's not the best but it's the only solution. Use this in your command line:

In your command line: ?1 ?s?d?u?l
left: Your dictionary
right: ?1

I put this at the end but I guess it can be anywhere: --rule-left="]"

Basically the rule deletes the last character and then you bruteforce every possible character at the end of your word.

#9
Thanks for that but it's still slow and with the size of wordlists, it's still no good.
#10
add the salts to a file and use it either on the left or the right side. of course this depends on the algo.