hp sws-2.2-1 (apache for OpenVMS)

[Neil Rieck]

Okay so I didn't see this one coming

I'm in the middle of migrating our business system from an AlphaServer-DS20e to an Itanium rx2800-i2

About 99% of the work so far is pretty-much a no-brainer (copy/compile/link/next) but my web services are unavailable on the new platform.

Our home page contains some jQuery which is employed to convert a linked-list into a sliding menu (ddsmooth) but most other pages at our site also make extensive use of jQuery and AJAX so if the home page has problems then so will everything else.

Weird browser behavior:
1) IE11 seemed to work properly
2) Firefox-39 rendered a page but no sliding menu; just a linked list
3) Chrome-44 didn't render anything at all but further examination with Developer Tool's (hit F12) showed a problem with incomplete chunked transfers

Rather than waste your time telling you all the things I did which did not work, let me say that CSWS-2.2-1 now requires text files to be stored in stream_lf format

Check out this official blurb:
http://support.hp.com/sg-en/document/c04739414

Has anyone else run into this? My notes show something similar with the release of CSWS-2.0 (2003-12-xx) which was fixed with CSWS-2.1 (2005-11-xx)

Neil Rieck
Kitchener / Waterloo / Cambridge,
Ontario, Canada.
http://www3.sympatico.ca/n.rieck/OpenVMS.html

[k...@kayceesoftware.com]

yes, had the same problem many moons ago and posted the issue and answer to multiple forums...

Also just heard back from VMS support in the past few weeks that a fix/patch has been created and they asked if i wanted it asap. I declined as i have just gotten back in the 'old' habit of not posting anything to our production server that was not stream-lf.

I was told in email from HP that the patch would be in the next formal release.

[Phillip Helbig (undress to reply)]

In article <b4c47c77-c809-4e9a...@googlegroups.com>, Neil

Rieck <n.r...@sympatico.ca> writes:

> Rather than waste your time telling you all the things I did which did
> not work, let me say that CSWS-2.2-1 now requires text files to be
> stored in stream_lf format

Another reason not to use it. TEXT FILES? On VMS?

[Jan-Erik Soderholm]

Have you tried the "interim image" that should fix this?

This is also a requirement with other OSS ported to VMS
such as Mercurial that is included in the Python kit. All
files managed by this CMS has to be in stream-LF before
loaded.

[Paul Sture]

On 2015-07-29, Neil Rieck <n.r...@sympatico.ca> wrote:
> Okay so I didn't see this one coming
>
> I'm in the middle of migrating our business system from an
> AlphaServer-DS20e to an Itanium rx2800-i2
> >
> About 99% of the work so far is pretty-much a no-brainer
> (copy/compile/link/next) but my web services are unavailable on the
> new platform.
> >
> Our home page contains some jQuery which is employed to convert a
> linked-list into a sliding menu (ddsmooth) but most other pages at our
> site also make extensive use of jQuery and AJAX so if the home page
> has problems then so will everything else.
>
> Weird browser behavior:
> 1) IE11 seemed to work properly
> 2) Firefox-39 rendered a page but no sliding menu; just a linked list
> 3) Chrome-44 didn't render anything at all but further examination
> with Developer Tool's (hit F12) showed a problem with incomplete
> chunked transfers
>
> Rather than waste your time telling you all the things I did which did
> not work, let me say that CSWS-2.2-1 now requires text files to be
> stored in stream_lf format
>
> Check out this official blurb:
> http://support.hp.com/sg-en/document/c04739414

Which includes "Most likely the issue also exists on OpenVMS Alpha, this
has not been tested/reported yet."

> Has anyone else run into this? My notes show something similar with
> the release of CSWS-2.0 (2003-12-xx) which was fixed with CSWS-2.1
> (2005-11-xx)

I ran into this in a subtly different way with a previous release (on
Alpha). It was something of a corner case* but I took the decision
to convert all the files to stream_lf to protect myself in the future
from the kind of thing you are seeing now.

The piece of DCL to convert files to stream_lf which shipped with
earlier releases of CSWS had disappeared from the then current release
so I had to fish it out from an earlier release.

* Google was pronouncing my robots,txt to be a dud and changing it to
stream_lf fixed the issue; IIRC the RFC for robots.txt states that
<CR><LF> line terminators are acceptable, so it appeared that CSWS was
serving it up incorrectly.

Side note: my testing of robots.txt suggested that Google's parsing of
robots.txt was "If we can find any excuse to ignore the entire file, we
will". They themselves recommend that if you really want to keep their
search engine out of a portion of your site you should password protect
it.

--
1972 - IBM begins development on its last tape drive (3480) ever because
of the declining cost of disk drives.

[terry+go...@tmk.com]

On Tuesday, July 28, 2015 at 9:42:44 PM UTC-4, Neil Rieck wrote:
> Check out this official blurb:
> http://support.hp.com/sg-en/document/c04739414

I suppose it is pointless to ask why HP (per that link) built Apache 2.0.65 on 1-JUN-2015, since it has been obsolete for over two years now (the last security fix from Apache that made it into 2.0 was for CVE-2013-1862). I suppose it is possible that they evaluated each subsequent security fix that went into 2.2 and 2.4 and backported the ones that affected 2.0, but somehow, given the neglect of these tools by HP, I rather doubt it.

I certainly wouldn't run 2.0.65 anywhere that it was exposed to "bad guys", either inside or outside a firewall.

[Neil Rieck]

On Wednesday, July 29, 2015 at 3:27:31 AM UTC-4, Jan-Erik Soderholm wrote:
[...snip...]

>
> Have you tried the "interim image" that should fix this?
>

[...snip...]

Not yet but we are considering it. Installing csws-2.2 then applying two updates will be a whole lot less work than tracking down all the code generating static pages (reports and such) every hour. (could involve checking/modifying 200+ programs)

We've also got people (quotes, sales support, etc.) uploading text files so would need to find all those trouble-spots then deal with them on an individual basis.

But if an official patch is available upon request then maybe I should try that approach first. Thanks for that tip.

Neil Rieck
Waterloo, Ontario, Canada.

[Neil Rieck]

I have no clue about their decision making process but my notes show Apache-2.0.47 first appearing with CSWS-2.0 (released in 2003-12) which changes to Apache-2.0.52 with CSWS-2.1 (released in 2005-11) then Apache-2.0.63 with CSWS-2.2 which is updated a few times. Perhaps some lower level bureaucrat just decided that CSWS-2.2-1 needed to be based upon Apache-2.0.something

Neil

[Neil Rieck]

While I'm still crabbing about this, here are are few more complaints about HP's "new" web site.

Here is the download page for CSWS
http://h71000.www7.hp.com/openvms/products/ips/apache/csws_download.html

The first link (documentation) takes you to a page of limited usefulness (IIRC, this is where you saw release notes as well as an installation guide; the newest release notes file for csws-2.2-1 in sys$help instruct the reader to visit "http://www.hp.com/products/openvms/securewebserver" where you see some familiar stuff but no iguide)

The second link (patches) behaves as expected

The first two download links work

The remainder download links take you you to a page of limited usefulness.

ps. people searching for older binaries may wish to visit "http://ftp.hp.com" then navigate to "/pub/openvms/apache"


Neil Rieck

[Neil Rieck]

Good news. I requested a patch from HP today via my OpenVMS support account. HP responded within an hour by placing the patch in a 24-hour drop box. I installed the patch ( a replacement version of APACHE$HTTPD_SHR.EXE ) and now everything works properly.

[Craig A. Berry]

On 7/29/15 6:37 PM, Neil Rieck wrote:
> Good news. I requested a patch from HP today via my OpenVMS support
> account. HP responded within an hour by placing the patch in a 24-hour
> drop box. I installed the patch ( a replacement version of
> APACHE$HTTPD_SHR.EXE ) and now everything works properly.

I'm glad there is a solution. It's hard to see it as completely good
news given that I'm pretty sure it's at least the second time the exact
same regression has occurred.

[Stephen Hoffman]

On 2015-07-29 09:37:02 +0000, Paul Sture said:

> On 2015-07-29, Neil Rieck <n.r...@sympatico.ca> wrote:
>>
>> Check out this official blurb:
>> http://support.hp.com/sg-en/document/c04739414
>
> Which includes "Most likely the issue also exists on OpenVMS Alpha,
> this has not been tested/reported yet."

I've verified that the CSWS bug exists on CSWS V2.2-1 on OpenVMS Alpha;
the content type is unexpectedly text/undefined. and the rendering goes
weird. Multiple browsers on multiple Microsoft Windows versions are
Not Happy with the various not-Stream_LF testing files, and to varying
degrees. With Stream_LF, they work.

I have also researched another oddity: no new access_log or error_log
entries. This though ErrorLog, TransferLog and LogLevel are all at
default settings, and the web server is up and running and serving
files, and SHOW DEVICE /FILES /NOSYSTEM SYS$SYSDEVICE: shows the
default APACHE$SPECIFIC:[LOGS]access_log. and error_log. files have
been opened by the APACHE$SWS process.

As it turns out, the log files are not flushed when written — even
overnight — and you have to restart the web server to see the access
log entries.

Related: <http://labs.hoffmanlabs.com/node/1912>


--
Pure Personal Opinion | HoffmanLabs LLC

[Jan-Erik Soderholm]

Isn't there anything in SWS like this option (in WASD)?

$ HTTPD /DO=LOG=FLUSH

[Steven Schweda]

> Isn't there anything in SWS like this option (in WASD)?
> $ HTTPD /DO=LOG=FLUSH

I have a self-re-submitting batch job which (hourly) does:

[...]
$!!! Apache 1.x:
$!!! @SYS$MANAGER:APACHE$CONFIG.COM FLUSH ! Bug causes lost log data.
$!!! @SYS$STARTUP:APACHE$STARTUP.COM graceful
$!
$! Apache 2.x:
$ HTTPD :== $ APACHE$COMMON:[000000]APACHE$HTTPD.EXE
$ httpd -k flush
[...]

I thought that it was working. I assume that this is/was
documented somewhere. (How else would I have known?) The
graceful restart also seems to be effective:

ALP $ type APACHE_GRACEFUL.COM
$ @ SYS$STARTUP:APACHE$STARTUP.COM GRACEFUL