26-Nov-2012

Working on WordPress upgrade
At this moment, all blogs are still on 2.6.5. The reason is simple> since WP 2.8 there have been internal changes that cause a loop in displaying the main pages.
Given the unexpected comment a few days ago, I think it better to get on to the latest version (l3.4.2) but I don’t want to interfere with this old version hat still works fine. So first, I created a new “blog” but still in 2.6.5: copied wp-config.php of an obsoleted blog and changed what’s needed, set up the system logicals to access it, and added the configuration in WASD_CONFIG_MAP. It created the blog and under 2.6.5, it all worked (including the hiccup in the admin pages).
Next, I installed WP 3.4.2, needed a bit of tweaking since unzip still fails to create a file including a dot in the name – not a big deal, there only 5 of them 🙂 – and set the logical for this test-blog to this version.
And now it didn’t work anymoe, as I was expecting: WASD runs into a loop because the mapping goes bazurk:
12:12:28.49  SERVICE   1659  0001  CONNECT      VIRTUAL  www.grootersnet.nl:80|
|12:12:28.49  REQUEST   4870  0001  REQUEST      GET  /wptest/INDEX.PHP/|
|12:12:28.49  MAPURL     1148  0001  MAPPING      PATH  /wptest/INDEX.PHP/|
...
Redirect:  /wptest/INDEX.PHP//index.php?
|12:12:28.50  REQUEST   1816  0001  REQUEST      REDIRECT  /wptest/INDEX.PHP//index.php?|
|12:12:28.50  SERVICE   1659  0001  CONNECT      VIRTUAL  www.grootersnet.nl:80|
|12:12:28.50  REQUEST   4870  0001  REQUEST      GET  /wptest/INDEX.PHP//index.php|
|12:12:28.50  MAPURL     1148  0001  MAPPING      PATH  /wptest/INDEX.PHP//index.php|
...
Redirect:  /wptest/INDEX.PHP//index.php/index.php?
|12:12:28.50  REQUEST   1816  0001  REQUEST      REDIRECT  /wptest/INDEX.PHP//index.php/index.php?|
|12:12:28.50  SERVICE   1659  0001  CONNECT      VIRTUAL  www.grootersnet.nl:80|
|12:12:28.50  REQUEST   4870  0001  REQUEST      GET  /wptest/INDEX.PHP//index.php/index.php|
|12:12:28.50  MAPURL     1148  0001  MAPPING      PATH  /wptest/INDEX.PHP//index.php/index.php|

It keeps adding /index.php until WASD finally concludes there is a loop somewehere….

So I removed all redirects in the configuration, but at that point, it doesn’t work either. It fails from the start:
|12:25:21.21 SERVICE 1659 0003 CONNECT VIRTUAL www.grootersnet.nl:80|
|12:25:21.22 REQUEST 4870 0003 REQUEST GET /wptest/INDEX.PHP/|
|12:25:21.22 MAPURL 1148 0003 MAPPING PATH /wptest/INDEX.PHP/|

even if I don’t specify anything but the minimal URL; index.Php should be added because that’s specified in WASD_CONFIG_GLOBAL as one of the starting documents. But it seems to be translated into a directory. And in that case, a 404 error (document not found) is to be expected, as is shown in WATCH output:
|12:25:21.23 ERROR 1539 0003 RESPONSE REQUEST:5965 (basic-only) 404(000) %X0001C04A (%RMS-E-DNF, directory not found) "Directory not found" "/wptest/INDEX.PHP/" "(no information)"|
|12:25:21.23 REQUEST 1099 0003 REQUEST STATUS 404 (Not Found) rx:648 tx:870 bytes 0.014648 seconds 103,635 B/s|

Being out of options for the moment, I asked on the WASD mailing list; I know some use WP as well and they have it running.
Of course, this test=blog has been disabled so you don’t need to try and access it :).

24-Nov-2012

Spam comment – how is that possible…
Anyone is free to comment, but you need to be logged in first…and before a commect will be visible, I need to approve.
But there were too many fake accounts added that I decided to disable creation of accounts by anyone, months ago.
So it was quite a surprise to be notified that some “Willy” at a site I don’t know in my system had pushed a comment. This is wat this mail told me:

Website : Willy (IP: 112.203.30.46 , 112.203.30.46.pldt.net)
URL : http://dripfeedbookmark.com/xxsmM3
Trackback excerpt:
.GOV LINKS BEAT PANDA!...

How's this blog doing? My site came back after Panda with this .gov link service...

Checked that site from the message: it’s cleary a abusive site; the address shows up in three blacklists…
So the commecnt has been marked as spam, and the address is banned from the network.

But there remains the question how this commecnt could have been pushed here anyway. Some investigation is needed…

21-Nov-2012

Helena updated
A few days ago already….
I had trouble with one non-critical program on Helena – the 16Gb Intel i7 system that is now my major workstation. Originally WIndows7Pro, I updated this machine to Windows8pro – because of the low pricetag :). It didn’t solve the problem (which is under investigation with the supplier) but offers some new insights in how systems are delivered nowadays. It takes a while to get used to it, but it dows the job, so far. Of course, I took precautions: a full backup of the system disk in case I want to revert – if possible. But so far, I’ve seen no real reason to do so. It’s different, that’s right. But, for me, workable.
Another advantage: IE 10 is winsockes aware. Propably not 100% but DClInABox works. And since it has been backproted to Windows7, I’ll install IE10 on these systems as well.

Webserver update
Mark Daniel published a new version of the WASD webserver a few days ago. And because I had still trouble using DCLInABox using Firefox15, and found a dissimilarity in the source compared to what the docs of this software states, he suggested me to upgrade.
I just dis, and all works – except the mail agent… but that’s merely a matter of file protection of the configuration file….Now to find out how to get around it….

As it turned out, it was indeed a matter of protection, but unexpected. In the update process, I enabled settting the site protecteion as it ought to be (no access unless allowed explicitly). Then I stopped the server with the shutdown proces. That removed SOYMAIL.EXE from memory; but I didn;t use the full startup, so SOYMAIL.EXE wasn’t reinstalled, and when executed on access, it lacked the privileges needed to access the configuration file and set logicals….
Once I had run the SOYMAIL startup – that will install the executable with proper privileges – that also worked again.
But it did not solve the DCLInABox problem with Firefox. Not even when running FF16 uncluding last updates: very much the same problem, where Chrome and IE10 are succesfull.
Next update might be Diana, moving up to VMS 8.4, and immediately apply the mandatory updates, and update 7 to get in line.

15-Nov-2012

Trouble in paradise
When I switched to fiber last year, I got a modem without a fixed IP address – wasn’t possible on that type, an besides: I would always get the same address on reboot – except for major network changes.
All went well until July, when a complete overhaul was executed – without notice, so I got a new address, so I had to contact the regristar of the domain to update DNS with the new addresses….A free upgrade from 50Mb to 100Mb, symetrical, for a one-time cost of EUR 10… Not bad. That address would not change, I was told. And when a major overhaul was expected, inclusing an address change, there would be a warning.
In the mean time, the ISP was bought by the big monopolist, and I got a letter that mentioned a chnage in administration. No action required.
Yesterday, about 01:20, connection was broken and restored within a minute. It happens now and than, and has no implications normally. But this time, the connection was restored with another address. No warning, either. So it was not before 09:00 that I found out NOT ANY connection could be made. Not even to the router – and setting up VPN failed as well.
That I got another address became clear when I looked into the situation that night. MAC addresses were still the same, even on the gateway, but the address was chnaged as well. As well as the DNS servers. Phoned the helpdesk, of course they couldn’t reverse the action, nor could they guarantee that it wouldn’t happen again; even worse, it could happen at any time from now on.. And a fixed address could no longer be obtained from them.
This is considered unacceptable, so I looked for another ISP. On the network, there is one more from the beginning: the big monopolist in fact, that offered less facilities for a higher price than the ISP I choose at the time. And for a few weeks, there is a third one – a subsidiary of this monopolist (but with their own standards). And as I found out that my current ISP is bought by that monopolist, there is actually no alternative any more for the facilities I require. But I will have all ports open – and a fixed IP address.
So I made the switch – for Internet, and by that, for HD television, because that requires Internet access as well. I vae a month notice period, so the definitive swtich will take place half December, it will mean a few days without Internet – I can live with that.
In the mean time, my sites weren’t reachable, since DNS contained the wrong addresses. So I asked my registrar to chnage addresses to the new one. This morning, these weren’t included yet, it took until noon today – after which I noticed I passed the wrong addresses, in all hectic of yesterday. That was repaired within an hour, and by and by, the changes propagated to the Internet. And at 16:30, this was the last site to be available again.
Now wait for the bill. EUR 10 enyway, perhaps twice – but that’s my fault. Hopefully, it will last until the switch of ISP (which will cost an additional EUR10 for that (final) change)

09-Nov-2012

Site changes in progress
The current homepage of this site is based on a set of routines I wrote years ago in Fortran77 on Vax, and included a program to test the function of the routines, and by that, it was used as a demonstration on how to use these routines. It uses the functions that VMS offers for memory management, and contained a number of MACRO routines; one in particular: that one is used to figue out which of teh parameters were actually passed: Fortran77 doesn’t really allow optional parameters and there was no other way to find out whether a parameter was actually passed, or not.
The Alpha architecture doesn’t contain the registers used in this macro so a one-to-one port wasn’t possible. Luckily, Fortran95 offers an alternative, and so it was easy to prot these routines to Alpha.
The demo-program was used to create a program that would create the HTML output.
Works great – and fast. But changing layout is quite a menace because the files are in a text library, and I still have to write a text file for the content. I have been looking for a way to change the text using a web-based interface but couldn’t find a good solution.
Then Mark Daniel – thw author of WASD – came with VWCMS – a VmsWasd based Content Management System. I have checked it in the past and it looks nice. It uses a javascript editor, that I cpould include in my program, for editing the main text. But VWCMS evolved, and it became quite feasable to use that. And so I started converting the home page into the format that VWCMS uses.
The basics are now done, and so I wanted to make it public. That means I have to adapt the WASD configuration as well, but how? So I asked, and I got an answer that does exactly what I need. I had to do some more adjustments, to be able to access the sub-sites (on OpenVMS , for instance), but the advantage outwages the extra effort: the configuration now holds clearly all parts that are accessable.

So that works fine, but I’m not completely satisfied with the homepage itself, there is some work to eb done before I can make it public. Until that time, the previous homepage will remain active.
Physical moves
The HSZ50-controller uses a battery to power the write-back cache; so no data is lost when deferred-write is interrupted due to a power outage. But that battery now fails, and write-through is disabled by the controller. And siince I could not actually use the shared-SCSI bus, there is no real need for this controller. For what I’ve heard, it slows down IO anyway: for the hardware, it’s a 10Mb/s connection but the disks are accessed with half the speed. The big advantaghe is that I can access multple cabinets – up to 6, though I have only 4 cables so I can access 4 at most….
So the question is: do I take it out of the chain? To be prepared, I made all disks in one cabinet “Transporable” so that cabinet can be taken out and be accessed directly, after I have created the contents using BACKUP/IMAGE. That is to be tested – I still have a PWS available for testing 🙂
If that works, I may swicth to basic SCSI-2 again….