• Welcome to OS2World OLD-STATIC-BACKUP Forum.
 

News:

This is an old OS2World backup forum for reference only. IT IS READ ONLY!!!

If you need help with OS/2 - eComStation visit http://www.os2world.com/forum

Main Menu
Menu

Show posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Show posts Menu

Topics - RobertM

#1
Hi, does anyone know how I can use REXX to get the following information? Inotherwords, a particular DLL Library I need? Something in the poorly documented RxSocks?

- TCP/IP traffic stats (similar to what the widget in xCenter/eCenter reports: current upload/download speeds and amounts)
- Active network/socket connections (ie: 47 connections to port 80)
- A tool that returns *VALID* data for the amount of available, free and used memory in the system (I'm getting all sorts of interesting results with the ones I am trying).

Best,
Rob
#2
Hello all,

I am hoping some of you have some recommendations for the best video card driver to use on my IBM eServer xSeries 346 servers.

The servers have an ATI RADEON 7000M video controller with 16 MB video memory. The following is what I have tried and experienced so far:

Default Drivers:
- OS/2 Warp Server for e-Business (CP2 PF) installs the GENGRADD non-accelerated driver
- The default installed driver allows higher resolutions and color depths (24bpp and 32bpp options to LCD max) (EDIT: to 1280x1024)
- Painfully slow redraws and such

SciTech SNAP:
- SNAP Full (v3.1.8 I believe) properly recognizes/identifies the video card, the amount of memory and the monitor
- Better performance than unaccelerated GENGRADD
- Will only allow up to 1024x768 (and indicates that is the highest the card supports)

Any thoughts?
Rob
#3
The following is part of the installation instructions to install OS/2 Warp Server 4 Advanced SMP on an IBM Netfinity 7000 M10:

- "All other partitions can be up to 512 GB using HPFS386, and 2047 MB for FAT partitions."
    http://www-947.ibm.com/support/entry/portal/docdisplay?lndocid=SCOD-3YYHJQ

I found that quite... amusing (and wish it were true...)

#4
Hello all, here are some tips for autostarting various applications on OS/2 and eCS.

I've run into a variety of interesting problems on eCS that I've finally tracked down. These tips may come in handy if you, like me, run a number of things that are REXX based.

Here's my scenario - various aspects of this may apply to you as well, if you've got certain REXX apps autostarting.


On my OS/2 server, I use xWorkplace and Styler/2. I have a timed startup folder that allows processing startup objects in order and with timed delays between them. eCS, by default, does not have the timed startup folder (HINT to Mensys, it NEEDS it). The server makes extensive use of various REXX function calls to external REXX DLLs, such as REXXUTIL (and almost a dozen others).

Besides setting up networking stuff, such as enabling the firewall and firewall rules, I *must* load the REXX DLLs on startup - the server is usually quite busy, and in order to provide more elegance in the REXX code and minimize processing time (by NOT checking whether the DLLs are already loaded each time the scripts are called by the web server), I load the DLLs at boot. In MOST cases, that may not be necessary, as the performance difference is minimal - but on a server that can often be called to process upwards of 100 requests a second, including calls to RxSQL, REXXUTIL and others, it does speed up request handling by not having each call to each script check to see if the DLLs are loaded - they simply assume they are.




Here's the fixes:

* Use an error handler that only is called on a script failure due to a DLL not loaded - then load the DLLs. That prevents repeated calls each script run to see if the DLLs are loaded - it only checks on failure - and should only do that once. (not the option I chose)
* Load all REXX DLLs on boot, preventing the need for extra code and thus streamlining the REXX code (a 100K REXX script/program may not seem like a lot - until over 100 of them are running at once) (I use THIS method)
* Have each script check to see if the DLLs are loaded when they are called - for the reasons noted above, I do not use this method.



Here's the problem with the method I use (for NON-xWorkplace systems)

(for xWP systems, simply enable and use the timed startup folder)
* Since startup items in the startup folders are not processed in a timed fashion, and simply try to load at the "same" time, if the REXX DLLs are not loaded first, other things in the startup folder fail - in my case, the web server control daemon (code for this is located in another thread - the code controls and monitors Apache and MySQL, ensuring they are running, autostarting on daemon failure, and doing timed shutdown & restarts of each to free memory pools). IF you have problems with certain REXX based things (or certain things that require REXX) failing to start properly from the startup folder, the solutions above and below may be helpful.



Thus, one can theoretically address this in the following fashions:

* Use one of the three methods outlined above (again, for the reasons noted, I prefer #2, but the lack of a timed startup folder causes issues)
* Install xWP and use its timed startup folder - that allows option #2 to work properly
* Call a REXX script in startup.cmd to auto-load the DLLs on boot. startup.cmd should be processed before the WPS stuff, thus giving items started by the WPS access to the already-loaded DLLs
* Install Rexx Autostart (by Herwig Bauernfeind with additions from Chuck McKinnis) and use that to control startup objects - though designed for various other issues (such as WPS hang on startup with certain objects in the startup folder), it is ideally suited for this issue as well.



A major advantage of properly addressing this issue, EVEN IF YOU DO NOT USE REXX, are the following:

* Let's say you are using MySQL and are protecting access to it via the firewall - you can ensure the firewall is configured and loaded BEFORE starting MySQLD, thus having NO time (even if it is only a few seconds) that MySQL is "exposed to the world"
* Other similar scenarios of course also apply.



So... there's the solution to a few different problems: (1) certain startup objects not starting due to DLLs not loaded before them, or (2) ensuring various objects are processed before others are loaded (ie: firewall before server daemons).

Hope that helps someone or anyone.


SIDE NOTE 01:
startup.cmd is NOT a REXX script. It is a BATCH FILE. You cannot simply dump REXX code into it unless you change the whole thing to a REXX script that processes the batch file code as batch file code. Thus, instead, you wish to call the REXX code as an application - and NOT simply dump the REXX code into the batch file. You must then also ensure that the REXX code exits properly so that the rest of the startup script is processed properly.




SIDE NOTE 02:
You can usually move firewall related stuff into startup.cmd as well - assuming you do not have other networking configuration stuff that loads afterwards which would change the necessary firewall rules AFTER the firewall is loaded - if that's the case, you can still load the firewall in startup.cmd, then load the new rule set during the WPS start via a script in the Startup FOLDER.




SIDE NOTE 03:
startup.cmd is NOT the startup FOLDER is NOT startup.cmd - the Startup FOLDER is a special WPS object. startup.cmd is a BATCH FILE that gets loaded on boot. MOST of you probably already know this, but every now and then, someone comes to the forum who is not even aware of the startup.cmd batch file - much less the differences between that and the startup folder - so I thought the differences were worth mentioning.




SIDE NOTE 04:
One may wonder why I simply do not call SysSleep in my Apache/MySQL routine - the reason is very simple actually - it is located in one of those DLLs that needs to be loaded first (namely REXXUTIL. Now, I could simply load all the DLLs in that control script - but other things require them as well, thus, other startup apps will still fail unless the server daemon control script happens to finish loading first.




-Rob
#5
Storage / Help: Recovering a volume
2011.10.21, 20:56:33
Hello all,

So, I've got an interesting scenario where a RAID failure has caused two volumes to be inaccessible. Both partitions (the information), one for each volume, are viewable in LVM and dfsee. The drive letters are semi-viewable in dfsee (ie: listed as "-d" and "-e" instead of "D:" and "E:"). Trying to reassign a drive letter via dfsee ends up with similar results. Inotherwords, using the LVM tools in dfsee and changing "D:" (which is how it is listed in the LVM tools) to "I:" results in the LVM information showing as "-i".

One partition is HPFS, while the other is JFS.

Any ideas would be greatly recommended.

Thanks,
Robert
#6
Hello all,

Here's an interesting question. I've got an old VIO app I wrote that I run full screen on an ancient OS/2 system I have. I am working on enabling it for a newer system.

Now... on to the problem. The app utilizes some very odd "mode" setting to set up the VIO screen size. (I think that size is something weird like 140x50). This is on an old Thinkpad system (TP600e and previously a TP600). If memory serves, I can use any number of columns and rows that is within the same limitation that a windows CmdLine session has (ie: COLxROW cant exceed xNumber (8,000 I believe)).

Interestingly, though the newer (various machines) and two old systems (the TP600s) are all running SNAP Full, the other systems I have seem to be limited to the standard original BIOS/DOS VIO modes (ie: the original 80x25, 80x43, 80x50) for full screen sessions.

Does anyone know if there is a way of bypassing those restrictions, or what's causing them?

Best,
Rob
#7
Applications / Apache & PHP Updates
2011.01.31, 21:58:25
Hello all,

I'm currently testing two things for Apache and/or PHP...
(1) sqlite for Apache (on SMP)
(2) APC caching engine

Hopefully I will have good results and Paul will feel confident enough with them for a general release. If you happen to be running a forum (such as the one we run) and have high amounts of traffic, APC supposedly does wonders.

And of course, thanks to Paul Smedley for his tireless work on these ports (don't forget to hit the donate button and drop him a couple bucks if you use and appreciate his efforts)!!!

Best,
Rob
#8
Hello all,

Does anyone know a way of forcing a reboot after a chkdsk? Say... after a crash and a large HPFS/HPFS386 drive being chkdsk'd? I have a couple systems that do just that, and one particular one that does not (which results in it hanging on various other driver loads - most notably NETWKSTA.200... folllowed by a trap, which spawns another reboot, which starts a chkdsk again, which results in errors loading NETWKSTA.200 and such, which is followed by a trap, which is followed by a reboot, which is followed by a chkdsk... you get the point.

I am not sure what the determining criteria is that's used by chkdsk.sys to decide on whether it will trigger a reboot or not; nor can I find any documentation that indicates a switch to always cause a reboot after the chkdsk - but I definitely have one system that needs an "always reboot" enabled.

Best,
Robert
#9
So, in trying to collect some more compatibility information on servers and SCSI cards and such, I keep running into various actual IBM server PDFs that claim the servers are compatible with OS/2 Warp Server for e-Business v5.0 - the most recent one (see link below) is dated 2004.

http://www.aibpc.com/Warp/235.pdf
Hardware Announcement from Jan 20, 2004 - years after WSeB 4.5 release.

I wish I knew what they knew that I don't...  ;)
#10
Applications / Dreamland BBS?
2010.12.08, 05:43:22
Hi,

Does anyone know who runs DreamlandBBS? I'm interested in mirroring that (and any other OS/2 related resource possible, before they all disappear).

Side note: the same applies to anyone else who has OS/2 information, files (drivers, apps, etc), articles, etc. I am willing to host and/or mirror them on one of our servers.

I am very interested in the information/articles/tips portion of any such stuff, so it can be converted to a format that will work on OS/2 World.

Best,
Robert
#11
Storage / Could this be ported to OS/2?
2010.12.03, 01:07:37
Hey all,

Could this be ported to OS/2?

http://en.wikipedia.org/wiki/IBM_General_Parallel_File_System

Just curious. It would be nice... for a number of reasons.
#12
Applications / Unexplained SYS1045 Errors
2010.10.31, 21:18:50
Hello all,

I've recently run into an error I don't think I've seen in over 15 years: SYS1045.

Some background:
  • This is on a system running eCS v1.2R

  • System has been running for over three years and (until this issue) only taken offline for the occasional cleaning. It's manage (between cleanings) 1 year, 11 months 2 days of continuous uptime... until recently

  • There have been no hardware changes, and no software changes (until this problem presented itself)

  • System is running: Web/2, FTP (Pete Moylan's), MySQL v5, NISTIME (for time sync), and wGET (to report status every few minutes)


The machine (at reboot, during normal operation, and after the error occurs), has only about 180MB of RAM in use.

Recently, every few weeks, everything that requires the commandline (including the commandline itself) will stop working, and no new sessions dependent on the commandline can be started. Attempting to start any cmdline will instead display:

SYS1045: The application program is not compatible with the version of the operating system being used.

Yes, that's correct... OS/2's very own CMD.EXE is no longer compatible with itself/the OS.

The full help message is kinda weird (HELP SYS1045):
EXPLANATION: The program requires a different version of the operating system.

ACTION: Use OS/2 Warp and restart the system or install the proper version of the application program on the system. Then retry the command.


Once restarted, everything works fine again - for about 3-5 weeks.

So... I'm baffled. I haven't seen this message since trying to run some very old OS/2 apps on a version of OS/2 that's significantly older.

Related notes:
- Closing all open programs does not alleviate the problem (as it does in "out of resource" issues)
- One can open various PM apps after this error has occurred. One simply cannot open/start anything that uses the cmdline (including the cmdline itself)
- Restarting the system always resolves this issue.
- ALL of the above listed apps have been running in the same fashion and on the same version for the better part of almost 4 years.
- Since the error started, we've upgraded FTPd, MySQL and Web/2... didnt help.

Any insight would be greatly appreciated,
Robert

#13
Hey all,

A while back I posted (probably on the old forum) that Giant (a supermarket chain in the mid-east US) had switched to OS/2 for their point of sale systems (back in 2000-2001).

Recently, I've found King Kullen (a very large local supermarket chain on Long Island) running an interesting operating system that I cannot yet place (never seen one reboot to look for an OS boot screen) that contains visual elements very reminiscent of OS/2 (such as the "3D" section containers/borders found around text/object groups in OS/2's settings notebooks and such, as well as similar looking controls).

While I am unsure about King Kullen, I do know, without a doubt, that Stop and Shop (formerly Grand Union) on Long Island is still using OS/2 for their Point of Sale systems. Not just that, but the one in West Islip, which is brand new (built this very year*), is among the ones running OS/2 (Warp v4) on the Point of Sale systems.

Just something interesting I saw and figured I would pass on.

Best,
Robert

* Was formerly a Grand Union - converted to a Stop and Shop a year or two ago. Ripped down to a hole in the ground. New building built and completed a few months back. OS/2 Point of Sale systems installed a few months ago. Thus, a supermarket will show as having existed there for a few decades, BUT the current one on that location isn't even a year old.
#14
Hi all,

The subject says it all. The specifics are as follows:

- eCS v1.2MR
- HPFS386 Boot Drive (has been since day two (literally, day two)
- Startup objects located in Startup folder
- SET RESTARTOBJECTS=STARTUPFOLDERSONLY, REBOOTONLY

I have removed "REBOOTONLY" and get the same results. My suspicions are this is an ini issue? Would this be a good guess?

Oddly, the system also ignores any JFS cache value over 64MB (only allocating 64MB regardless of the setting) even though the system has 2GB of physical RAM.

Thanks,
Rob
#15
Hi all,

Can someone outline the best way of upgrading eCS V1.2MR, specifically:

  • Kernel
  • JFS
  • DOSCALL1.DLL

I had thought there was a package out there that upgraded the kernel and DOSCALL1 as needed, but cant find it.

My current DOSCALL1 is dated 10-10-2003 137,469 bytes - BLDLEVEL cannot report a version number for it for some reason.

Thanks,
Rob
#16
Hey all,

I've been reading more and more about various vendors who were stupid enough to believe Microsoft's promises of Windows 7 being able to run on the newest tablet similar to HP's Slate and others.

I've been thinking, with the extensibility of the WPS, capabilities of Firefox, upcoming Flash, and eCS's minimal resource requirements, wouldnt it make a good option?

Yes, I know there are other things that would need to be addressed (WPS extensions, Flash being released, an updated Java, multitouch drivers, Telco wireless support (ie: 3G, etc)) - but to me it seems a great new market without a tremendous amount of work - assuming that work to get it running on the CPUs wasnt that extensive.

It seems more hardware manufacturers are looking for Windows 7 alternatives for Atom and ARM based tablets... What do you all think?
#17
Storage / JFS eating CPU resources?
2010.04.28, 23:36:59
Hello all,

I have now noticed this on two of my machines. It seems JFS really really does not like doing more than one thing at a time. I have a 64MB cache configured on one, and a 32MB cache configured on the other. The first is definitely the latest version of JFS I could find (and latest SMP kernel). One is U160 SCSI, the other is a combination of SATA and IDE.

The machines:
- Netfinity 7000 M10 4GB RAM, U160 SCSI RAID configured to RAID 1E (striped and mirrored) running WSeB CP2 PF on FOUR CPUs.
- Intel 845 Chipset board with one IDE and one SATA drive and relatively recent DANI drivers running eCS v1.2MR on single 2.8GHz P4

The problem:
When transferring decent sized files (FTP, disk to disk copy, etc), it seems to eat a lot more CPU than it should. HPFS386 does not seem to have this problem (especially at the TWO concurrent FTP transfers or copies I limit these actions to).

I've attached a screenshot of the CPU status bar/etc. This baffles me. The CPUs should be at between .1% to 5% - instead they are hitting up to 25%

Upload or download doesnt seem to matter too much (upload to the server seems a little worse).

Best,
Robert
#18
Hello all,

I am looking for information on the PROCESSES= statement for the config.sys file in regards to WSeB and eComStation. The info I have found so far is very vague.

This is for a server that is generally under high load running Apache. Any help would be greatly appreciated.

Best,
Robert
#19
Hey all,

This may be a silly question... it may even be one I asked (or even answered) in the past... so apologies if it's been asked and answered before (couldnt find it on the boards).

I am looking for a REXX Extension DLL that provides HTTPS support in a similar fashion that curl does. As a matter of fact, REXXCurl would be wonderful, if I could find a compiled version of it.

I am writing a small script that needs to do such things. Currently, it works great using Curl, but thus requires RxQueueing the output or sending the output to a file to be read. I'd prefer a better option.

I am presuming wGet will also work fine - but also seems to lack a REXX interface DLL.

Any suggestions are greatly appreciated.

Thanks,
Robert
#20
Hello all,

We are looking for a programmer who has some experience with:

(a) building REXX DLLs
(b) porting *nix code

Namely, we are interested in:

(1)


Currently, I have an Amazon AWS account that requires key authentication - for which I use a PHP function included in your builds.

It uses the hash_hmac() function. All works perfectly and as expected under Apache and PHP. I still have the need to run a few Lotus Domino Go WebServers running scripting via REXX though, and currently, my workaround for those is to send the request to one of the Apache servers, and have it pass back the URL string with the hashed value...

Preferably, I would love to be able to access that function via REXX directly on the LDGW boxes instead of the less elegant method I am currently using.

Is it at all possible to compile a version of that function that are REXX callable?



(2)


We are hoping to obtain the same functionality out of the Apache PDF plugin where it can be REXX callable with the appropriate HTML passed to it so it can then output the PDF version.



These can be individually tackled by different people - or both done by the same person. Both DLLs needed have already been ported to OS/2 by Paul Smedley, so I suspect the porting portion of the work should be pretty easy... which just leaves adding the REXX entry and exit points...

Any takers? Offers?

Best,
Robert