Month: February 2008

MySQL, again
It continues to cause problems.
Twice, in a short period, it crashed with the same error:

080217 19:37:09 InnoDB: Error: the OS said file flush did not succeed
080217 19:37:09 InnoDB: Operating system error number 12 in a file operation.
InnoDB: Error number 12 means 'not enough core'.
InnoDB: Some operating system error numbers are described at
InnoDB: http://dev.mysql.com/doc/mysql/en/Operating_System_error_codes.html
InnoDB: File operation call: 'flush'.
InnoDB: Cannot continue operation.

both when updating the bootcamp 2008 blog (quite a lot of them: I was looking for a proper layout, updated existing entries, and added a few new ones.). I raised the BYTLM and PAGEFLQT parameters for the server (both 150 %) , perhaps that will help. I’m even thinking of installing MySQL 5..1, which seems to be more stable. But the question remnains if Diana is big enough (256Mb) tho cope with it. (Or is MysQL so rigid it cannot cope with paging?)

Actually, this is all. Except that there are a number of patched to be downloaded and installed, and the forum needs to be cleaned from bogus users (when done, theer were about 40 entries to be deleted, But with the plugin, it’s a matter of a few mouseclicks)

PC Cleanup
This weekend was used to do some cleanup on Aphrodite – burn a lot of not-too-urgent, read-only data to DVD and CD in order to free some room for new data. Part of what’s currently on the C drive shouldn’t be there in the first place to it will be moved elsewhere, but in order to do so, some space had to be freed – and that meant archiving files to read-only storage. A lot of older files – the 2004, 2005 and 2006 bootcamp files, for instance – have been moved off disk, and all photographs that were processed to be published as well. Stored downloaded software went to DVD as well as finished projects.
On the VMS front, a minor glitch in backing up the webs has been solved: there was a concurrency issue where creation of the operatorlog index purged files that were already tracked down by BACKUP, doi9ng the backup of the webs – and that caused BACKUP to fail because one of the files to backup wasn’t found. So I disabled purge when webs are backed up, just to prevent these problems. A purge will happen just one day later.
Content program
The initialisation module has been changed to do all init in one call: The web itself, and all blogs under it. I also created the CGI-module, based on a module I created (in Dibol) in some project a year ago. It compiles well, now it’s time tio test it. Since that requires web access, logging should be enabled – meaning yet another addition in the library, and changes to the initialisation module. Both are considered VERY minor….
It’s another matter to split the signalling module in the library, in order to prevebnt the requirement of FDV in linking. It will be just a warning but I consider it ‘not done’ to require the FMS runtime to be available when you don’t need it.
After that; I’ll first do some documentation on the library modules so the files can be published.
Blogs
In the mean time, I’m working on updating the blogs to match WordPress 2.3, and that means I’ll have to do something with themes as well. The first blog I’ll do is the Bootcamp2008 one. The others will stay on 2.2 until my native program is finished. Unless I find a nice theme I can use.

There was an issue with the test blogs after I changed the operator web to SSL: I have the WordPress software accessed directly fro testing reasons: the blogs still tried non-SSL and therefore none of the related links: CSS, Admin entry… worked. Getting the URL in the paste-buffer, pasting in into the URL enry of the browsen and chnaging the protocol to https:, I was able to change the specification from within the blogs. Now they show up like they should and the admin page works again; Links within the posts however still need an update (should also be https:) so images will show up as well.

The programming continues
The program to handle dynamic contents is progeressing slowly – there are just a few hours that I can work on it so it’s just small steps that can be taken. At the moment, the basic IO is done, as well as basic initiazation for the program. Next is retrieving the webserver data that is rquired to get the right data. I still have to see what difference to use. Both the web’s domain as well as URL are possible as webname. For instance, either “www” of “sysblog” could be valid. This should propably be part of the initiatization module; reading the initialization file however is correct at this moment. I’ve choosen for a fixed file, specified by a logical, and if this is not defined, use a default name and default location.
Blog matters
Just found that, once again, the blog’s theme seemed ‘lost’ and so I had to re-enable the right one. Someting that happens sometimes…
An other thing that is found not working is VMS help, it’s likely to be caused by a missing logical. Until I solved that issue, the link has been disabled.
UPDATE
I don’t know what was wrong, but the link was simply correct. It might have been the proxy server at the site where I encountered the problem. So I enabled it again.
This morning, vamp.issinoho.com was online again, but tonight it seemed off once again. I’ll keep it offline (and ask Iain what’s wrong 🙂 )
Forum cleanup
Some bots lie to add “users” to the forum but the new software won’t allow their drug- and porn-related sites (and some plain malware-spreading ones) be shown. Removing them is now a matter of just a few clicks – and gone they are. Today, a two-week harvest of 49 bogus members.
Mail statistics
Total messages   : 3699
DNS Blacklisted  : 2356 = 64%
Antirelay        :  379 = 10%
Scanned by PMAS  :  964 = 26%
    Discarded    :  320 = 33% of processed, 8.5% of total
    Quarantained :  257 = 27% of processed, 7 of total
    Valid        :  117 = 12% of processed, 3% of total

That means that 3/4 of all mail is simply discarded (I never see it) and almost 90% of what passes the filter is to be ignored (except when it’s a false positive – there were a few last month).
Last month, 5 messages slipped the filter and were filtered off by SMTP itself.