Maintenance
The maintenance job has started, did some work – but not entirely correct, mainly because of specifications that weren’t entirely correct, of missed a $ SET DEFFAULT and a typo, but since deletions have been commented out, is was not a really big issue, and the errors have now been corrected, and, as far as possible, tested. Waiting for the next automated run….
Mail statistics: A lot of relay attempts!
What did work, anyway, was mailing the mail statistics:
PMAS statistics for June
Total messages : 4975 = 100.0 o/o
DNS Blacklisted : 2016 = 40.5 o/o (Files: 30)
Relay attempts : 1888 = 37.9 o/o (Files: 26)
Processed by PMAS : 1071 = 21.5 o/o (Files: 29)
Discarded : 164 = 15.3 o/o (processed), 3.2 o/o (all)
Quarantained : 334 = 31.1 o/o (processed), 6.7 o/o (all)
Delivered : 573 = 53.5 o/o (processed), 11.5 o/o (all)
Again a high number of relay attempts: over 700 on June 19th and 25th. On the 19th, the offending address was 118.249.218.246, the supposed senders were jxzhn@21cn.com (180), csj-lx@sohu.com (164), andly88888@163.com (188) and haoyunwp@sina.com (188), they cycle after a number of attempts. On 25th, the offending address was 113.97.110.230, sent by either asdjkl831266@126.com (348) or asdjkl831266@163.com (348). Far less attempts have been found on June 5th (81, from address 219.132.25.98, by hioaghoghl@126.com (47) and iaghghlk@126.com (34)) and 7th (291, from address 222.247.120.209, by jacklywangpeng@126.com (70), 2006peixunbu@21cn.com (80), haoyunwp@sina.com (64) and wphaoyundao@sohu.com (77)). And I’ve seen that July will hold even bigger numbers: on 03-jul-2010, there have been over 2000 attempts from address 121.13.12.72, by sender “smtp100@sina.com” , every second between 9:30 and 10:21…
All addresses are located in China, 219.132.25.98 and 121.13.12.72 are likely to be broadband addresses and could be forged (according the query result), the other ones link to CHINANET-BACKBONE. But all tried to access (non-existent) accounts at my ISP via my server. But it doesn’t accept relay – my ISP tries on regular intervals, and I see them refused as well.
Irene ‘renewed’
There are continously problems with DNS translations; though all settings seem Ok, and Internet access is never a problem, accessing local systems is notoriously problematic. Using the standard internal domain (intra.grootersnet.nl) any non-Windows system cannot be found, but without, there is no problem at all.
But this time, I had enough. I installed Ubuntu on the machine as the primary OS, and that solved this problem. The general users on that system (my wife and son) usually just use the web-browser, and my wife will occasionely use email but never store a message. They’ll have to learn new interfaces, but the differences are minor.
Main advantage si that my wife now has most of the menus and texts in Dutch – like she prefers, where my son and I have no problem with the English menus.
One problem, though, still to be solved: I changed my wife’s password for login, but her keyring-password hasn’t changed, and I’m looking on how to do that.
New web- and PHP getting on
Last weekend, a new release of WASD has been published, which fixes a few issues I ran into. So that was the first thing to do: install the new server. Normally, this is just running a proceduree, but this time, there was something wrong: the OpenSSL package that I use was not correctly installed, but after that, installation went fine – except that a new file (required when updating from 9.3) was missing. But since this installation is a fresh one, it didn’t effect me.
The server runs fine now.
PHP is another matter.
It took some reading. mailing, deduction and trial-and-error, but I finally got PHPMydmin and WordPress running. Not entirely 100%, but at least I found the problem: both applications had problems including files – couldn’t locate them, and run into an error. A closer look on the PHPMYADMIN log showed a possible reason: ./libraries/common.inc.php – there are more than one dot in the file…
Copied the file to be “common_inc.php” and changed to code accordingly: behold, the error was gone, but the next multi-dot file caused the same error.
Wordpress ran into a similar error.
Now I looked at the HP PHP documentation, and there is stated that setting parse-style to EXTENDED – what’s done by WASD if an ODS-5 file system is encountered – is not sufficient: You need to set the following logicals:
$ DEFINE /NoLog DECC$EFS_CASE_PRESERVE ENABLED
$ DEFINE /NoLog DECC$EFS_CASE_SPECIAL ENABLED
$ DEFINE /NoLog DECC$EFS_CHARSET ENABLED
$ DEFINE /NoLog DECC$FILE_SHARING ENABLED
To try this, I set these logicals /SYSTEM and now PHPMyAdmin got into the login screen, and WordPress asked the credentials to create the database.
So I had to change the configuration to use PHPWASD.COM as a wrapper – that sets these logicals before invoking PHPWASD.EXE. It made a difference. PHPWASD.EXE makes an attempt to set these as well, but it seems not enough.
So there is some progress, but there are still a number of things to check: PHPMyAdmin cannot connect to the database and gave a “heap corrupted” message, and WordPress ran into ACCVIO errors. But after $ DEFINE/SYSTEM USE_ZEND_ALLOC 0 both seem to be gone.
But still, phpmyadmin cannot access the database, and wordpress now runs into a loop: accessing “/index.php” redirects to “/INDEX.PHP” – and I get a “page not found” error…