After updating samba from 3.0.1 to 3.0.2a ntlm_auth in NTLMSSP squid 2.5 helper mode says frequently (but not every time): [2004/03/17 21:26:38, 1] libsmb/ntlmssp.c:ntlmssp_server_auth(574) ntlmssp_server_auth: failed to parse NTLMSSP: [2004/03/17 21:26:38, 2] lib/util.c:dump_data(1830) [000] 4E 54 4C 4D 53 53 50 00 03 00 00 00 18 00 18 00 NTLMSSP. ........ [010] 55 00 00 00 18 00 18 00 6D 00 00 00 06 00 06 00 U....... m....... [020] 40 00 00 00 07 00 07 00 46 00 00 00 08 00 08 00 @....... F....... [030] 4D 00 00 00 00 00 00 00 85 00 00 00 06 02 00 00 M....... ........ [040] 53 49 42 52 4F 4E 4B 4F 53 54 59 41 4E 48 45 4C SIBRONKO STYANHEL [050] 50 44 45 53 4B 54 46 87 0B 9D 3A B4 1A 96 CD A5 PDESKTF. ..:..... [060] 02 02 79 82 EC 87 F7 62 C4 E0 5E B2 83 8F BD 42 ..y....b ..^....B [070] 78 BA A9 10 09 10 72 CE 9E 3C 2B A2 D6 2C 25 C6 x.....r. .<+..,%. [080] 94 AD 73 74 EE ..st. The PC is Win2K PRO workstation with SP4 and all latest Windows Update critical patches installed. Any ideas?
Just a note to say I've seen this, and that I think it's probably real. I'll copy out your log extract, and see what I can make break. NTLMSSP is fun! ;-)
*** Bug 1475 has been marked as a duplicate of this bug. ***
Andrew, is this fixed ?
Still seeing this with Samba 3.0.9 & Squid 2.5-STABLE7. Debug from cache.log reports Got 'KK <Truncated>' from squid (length: 139). Got user=[<User>] domain=[<Domain>] workstation=[] len1=24 len2=0 From what I can see this doesn't look to be related to Samba, I saw there was a bug fixed on Stable7 related to truncated NTLM auth, not sure how to go about pinpointing it back to Squid or maybe a Windows error. Any ideas?
andrew, please update this somehow. Still working on it. wont fix. etc... thanks.
I don't think comment #4 belongs here - that shows the 'auth' packet being parsed correctly, and so is some other error. The initial report my be a truncated blob, rather than anything more intersting. (ie, try again with current Squid 2.5 STABLE8).
Original reporter did not indicate that authenication failed, so I tend to think that the issue is one of cosmetics. I've removed a bogus printing of the 'NTLMSSP parse failed' line from current SVN, and as such I think this is fixed.
sorry for the same, cleaning up the database to prevent unecessary reopens of bugs.