MySQL crashed – again
Just after posting the Babelfish interpretation of “Flushing Cache”, MySQL server went down. This time, it took a few hours to get up again – because I had no access to the systems. It’s been pretty much the same: value 12, “not enough core”, this time on read.
Reverse changes
I took the chance to reverse a few changes made in favout of the IDE server for Distributer Detbenas and WEBES and lowered some of the system parameters that wer increased as their requirement, but since RdB is to be installed some day, I kept an eye on the requirements. The only one needed a permanent increase comapred to the original setings, is CHANNELCNT – set to 2050, just less than twice the original amount. The rest have been restored to the original.
Next I rebooted Diana (some of them are not dynamic) but something strange happened: It looks like some command procedure has “SET VERIFY” in it and the whole listing was sped out on the screen. MySQL was started but the remains of the procedure seemed to be bypassed. Running it separately revealed no troubles at all, so what happened here, I don’t know. The log didn’t shbow anything either….
But the system is now working again.
Interpretations…
How do you translate “Flushes Cache” in Dutch:
In more or less ordinary English:
The flood’s secret portal to memory
(Thanks to my collegue Piet Timmers)
Paypal phishing attempt blocked
The phishing attemps are now blocked by the anti-spam gateway so they do no longer arrive in any of my mailboxes. Which, of course, is what it is the intended use, plus it allows a closer look to the message code without having the message actually delivered.
This one came in a few days ago:
<p><b><font face="Verdana" size="2">You are required to upgrade your PayPal
Account by subscribing to our New Security Center.</font></b></p>
<p><font face="Verdana" size="2">Please <b> <a href="http://202.67.156.66/www.paypal.com/cgi-bin/webscrcmd=_login-run/update.php">click here</a></b> in order to upgrade your PayPal account.</font></p>
<p><font face="Verdana" size="2">If you not perform the update now, your account will be placed on hold. On hold accounts can still send money, but they cannot withdraw or receive funds.</font></p>
Mind the hyperlink-address:
http://202.67.156.66/www.paypal.com/cgi-bin/webscrcmd=_login-run/update.php
This is NOT a paypal address.
PMAS signalled this – as is shown in the message header:
Received: from unknown ([72.54.216.109] EXTERNAL) (EHLO mail.iei-web.net) by
xxxxxxxxxxxxxxxxxxxx ([192.168.0.200]) (PreciseMail V3.0); Sun, 07 Oct
2007 06:41:42 +0100
Received: from User [62.14.249.101] by iei-web.net with ESMTP (SMTPD-9.10) id
A0F40294; Sat, 06 Oct 2007 23:39:00 -0600
Reply-To: <member_service@paypalsecurity.com>
From: "PayPal Inc."<member_service@paypalsecurity.com>
Subject: New Paypal Security Center: Update Your Account
Date: Sun, 7 Oct 2007 07:40:01 +0200
MIME-Version: 1.0
Content-Type: text/html; charset="Windows-1251"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 6.00.2600.0000
X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
Message-Id: <200710062339882.SM03048@User>
What are the findings:
X-PMAS-External: unknown [72.54.216.109] (EHLO mail.iei-web.net)
X-PMAS-Software: PreciseMail V3.0 [071006] (diana.GROOTERSNET.NL)
X-PMAS-DYN_URI-OK_URL: Dynamic URI check: OK URL (0.000)
X-PMAS-VMF-OK: Envelope FROM: check: Source accepts mail for address (0.000)
Quite well so far – except the “unknown” external address.
But now the problems show up:
X-PMAS-HDR-MISSING_HEADERS: Missing To: header (1.035)
X-PMAS-HDR-CTYPE_JUST_HTML: HTML-only mail, with no text version (1.500)
X-PMAS-HDR-RCVD_FROM_UNKNOWN: Message received from host without DNS entry (4.000)
X-PMAS-HDR-NO_SPACE_FROM: From: header is poorly formatted (no space) (5.000)
X-PMAS-URI-NORMAL_HTTP_TO_IP: Uses a dotted-decimal IP address in URL (0.942)
X-PMAS-URI-IP_LINK_PLUS: Dotted-decimal IP address followed by CGI (0.708)
X-PMAS-META-FORGED_OUTLOOK_HTML: Outlook can't send HTML message only (1.101)
X-PMAS-META-MISSING_BODY_TAG: Message has </BODY> tag, but no <BODY> tag (3.000)
X-PMAS-META-MISSING_HTML_TAG: Message has </HTML> tag, but no <HTML> tag (3.000)
X-PMAS-META-FORGED_OUTLOOK_TAGS: Outlook can't send HTML in this format (5.000)
You learn something new every day
X-PMAS-META-NO_HTML_BEGIN: Message has </html> but not <html> (3.500)
X-PMAS-META-PHISHING_01: Message is a phishing scam (50.000)
X-PMAS-META-PHISHING_03: Message appears to be a PayPal phishing scam (20.000)
X-PMAS-META-LAME_PAYPAL_SCAM: Claims to be from PayPal, but no PayPal URIs (20.000)
I thought so 🙂
X-PMAS-META-CLICK_BELOW: Asks you to click below (0.727)
X-PMAS-META-BLIND_DATE3: Blind date spam (3) (20.000)
X-PMAS-Final-Score: 139.513
X-PMAS-Spam-Level: ********************+
X-PMAS-Spam: Yes
Apart from what is unusual in Paypal: no addressing header (should use your Paypal name).
Updating FORTRAN code
I’m in the process of porting code written on VAX in FORTRAN77 – and some routines have optional parameters. FORTRAN 77 cannot handle these, but on VAX, it’s easy to retrieve them using a small MACRO module. Not so on Alpha, and Itanium would be even more troublesome – if possible at all due to the architectural differences and processor technology. (See earlier thredas in this catergory).
So I have to choose: re-write the code in another language, or update it to a newer FORTRAN version that allows checking of arguments, as Hoff suggested.
I choose the latter – it proved a far less troublesome task than anticipated.
Consider a routine A, with 4 arguments K, L M and N. K being required and L, M and N optional; however, any of these could be present.
This is the basic FORTRAN 77 code:
SUBROUTINE A (K, L, M, N)
C arguments
INTEGER*4 K
INTEGER*4 L (*)
CHARACTER*(*) M
LOGICAL*4 N
C Local variables
INTEGER*4 S, nmbr, mask
C Routine used to see whether args are present
C Bits in Mask will be set when present
INTEGER*4 GETARGS
C main code
S = GETARGS (Nmbr, Mask)
...
IF (BTEST (Mask,2)) THEN
C
C Parameter L is present
C
ENDIF
....
This routine can be called as:
F = 1
G(1) = 1
G(2) = 1
G(3) = 1
H = "This can be a text of arbitrary length"
I = 0
CALL A (F,G,H,I)
CALL A (F,,H) ! so L and N are missing
Updating the routine to FORTRAN 95 so the optional parameters can be handled, are really minimal: add a line that specifies which arguments are optional, remove the call of the GETARG routine, and use PRESENT (arguments) in stead of BTEST(Mask, bit).
The FORTRAN95 code look like this:
SUBROUTINE A (K, L, M, N)
C arguments
INTEGER*4 K
INTEGER*4 L (*)
CHARACTER*(*) M
LOGICAL*4 N
OPTIONAL :: L, M, N
C Local variables
INTEGER*4 S
C main code
...
IF (PRESENT (L)) THEN
C
C Parameter L is present
C
ENDIF
....
To use this, the routine calling this subroutine must specify the interface. If a routine is heavily used, it’s worthwhile to create an INCLUDE file containing the interface:
INTERFACE
SUBROUTINE A (K, L, M, N)
C arguments
INTEGER*4 K
INTEGER*4 L (*)
CHARACTER*(*) M
LOGICAL*4 N
OPTIONAL :: L, M, N
END SUBROUTINE
END INTERFACE
Add this into calling routines:
INCLUDE 'A_IF.INC'
...
F = 1
G(1) = 1
G(2) = 1
G(3) = 1
H = "This can be a text of arbitrary length"
I = 0
CALL A (F,G,H,I)
CALL A (F,,H) ! so L and N are missing
This is still to be done, but it looks good!
I ran into one problem still to cope with:
CALL B (%VAL (Args))
does not compile: %VAL is out of context here.
The routine in which this code occurs is rather basic – and the mechanism is used heavily in calling routines, where the addresses of allocated memory are passed….
UPDATE
It turned out pretty straight forward. This particular reference turned out to be a parameter to a routine that could easily be bypassed by assigning the right value to a separate variable and use that one instead.
Well, all modules have been built now – except for the ones requiring a missing file but I don’t need that one in due time – and the libraries are created. Next is translating the macro containing the translation vectors into an option file, and vreate the shared image. Afther that, I can start creating the drivers to test it all
14-Oct-2007
Persephone updated
Last week, a new version of Personal Alpha was released. Still a DEC3000 but enhanced to 128 Mb internal memory, and the full SRM console. This has been installed over the previous version, with some trouble: I should have de-installed it before installing the new one. This wasn’t mentioned! But after I did, the first invocation installed it and got it running. (Part of the issue was the relation between the .EMU extension and the assiciated program. Once that was solved, it’s all done)
PMAS crashed
One of the PMAS workers has crashed – I still have to check the logs ;(
%%%%%%%%%%% OPCOM 13-OCT-2007 12:02:51.30 %%%%%%%%%%%
Message from user SYSTEM on DIANA
%PTSMTP-E-WORKERDIED, worker PTSMTP 0001 (20200150) terminated unexpectedly
-SYSTEM-F-ACCVIO, access violation, reason mask=00, virtual address=000000000000FE79, PC=FFFFFFFF81F6CFC8, PS=0000001B
-PTSMTP-I-WORKERCONN, while processing connection from 86.57.187.251,2132
Perhaps because theer was something illegal from this site?
The main program was not effected so mail just went on. It’s just this connection.