I'm chasing an irritating - not disabling - ghost regarding intermittent crashing of Sesame 2.5.3.

Small network of 8 computers, all running WinXP with 2gb RAM (except one new laptop running Win7-64bit).  Lots of transactions (sales orders, purchase orders, inventory received, printing of orders and labels, invoices) all day long.  Lots of emails generated via @sendmail as a part of some measure of the transactions (alerts sent to customers, alerts sent to employees for various things).  And I use a LOT of XResult commands in my programming.

Since upgrading to 2.5.3 from 2.5.2, there is a disconcerting increase of Sesame crashing - both the client, and sometimes the server.  We might go a few days without any crashes, and then have 4 or 5 over a two day period.  Some workstations are worse offenders than others.

The most common circumstances seem to be 1) @sendmail, and 2) printing labels to a Brother QL580 label printer. 

Here are the variables -
1) if some specific workstations are hanging up while attempting to execute code that includes a @sendmail, then rebooting the workstation will 'most' of the time allow the workstation to then perform that Sesame code successfully again.  However, at least one of the workstations seems incapable of performing a @sendmail, even after fresh reboot (2gb ram).
2) rebooting the WinXP workstation that functions as the stand alone Sesame server every week or so seems to help
3)  the Win7 laptop, with 8gb RAM (upon which all the Sesame development is done) statistically NEVER (once in a very very long blue moon) has any problem running any aspect of the program.

So ...
   - do I have a Sesame problem? 
   - Do I have an 'aging workstations' problem and need to upgrade my workstations to faster processors, more RAM, and Win7?

Comments:  I've got two 'quick interim fixes' going -

"Avoidance" fix - for certain UserId's, I've commented out some of the @sendmail code, and

"Muscle" fix - I've got a Win7-64 bit 8GB workstation (used) on the way, to insert in place of the most frequently offending XP workstation and see if its muscle power  resolves the issues.

Any advice or suggestions for me?

Mark,

Agree.   But while "software diagnostics and repair" is the default 'fix' on the 50th 'brain' floor where you think and work, down in the engine room amidst the coal and the shovels, I'm going to try the ham-fisted 'muscle it to death' approach first, as I find myself with a shrinking time/attention span window these days.

I never have any trouble with any Sesame thing between my  new laptop with big RAM, fast CPU, and Win7-64.  So, I'll try sticking a desktop with the same type specs on the network and see how it goes.

Sendmail - mostly the error code is either a -1 or a -4.

I'm convinced that Brother's label printer printer driver for Win7 is much better then the one for WinXP ... I'll see if the proof is in the pudding.

Many thanks.

Ray,

the -1 return code:  this usually is the case when I'm running Sesame in standalone, doing some development or some such on my laptop.  The @Sendmail does NOT like stand alone mode or when I am a client to the office Sesame server through the Internet (at least as I have it set up), but it runs like a champ if my laptop is a physical client on my network at the office.


The -4 return code:   this is a recurring error from one computer most of the time (maybe it's all the time on that computer ... I should disable the security software and test it a little bit), and a problem on several other computers intermittently, although if it happens and then we reboot that workstation, it almost always allows Sesame to operate error-free for awhile.  But if I get an -4 error on one of these slaggard workstations, and then go run that command button from my newer Win7-64 laptop, it always 'just works'.  The 'To, CC, BCC addresses are hard coded in several of these buttons  to send alerts to several of my employees throughout the day on various things (although I am reprogramming the hard coding, as I 'get to it', to allow for more flexibility).

On the crashing issue, a couple of the older computers (6 years old?  2gb ram, WinXP) can experience client Sesame crashes several times a day, but then go a few days without issues (usually, as I've mentioned, when printing labels - I think Sesame get's irritated at having to wait for the slow speeds of the label printer and the spooler ... one reason I'm anxious to put a desktop on the network running Win7 with 8gb ram and see if that problem disappears from that network node).  But the Sesame server engine will crash sporadically maybe once a week (it's hard to find a pattern - it might happen a couple of times in one day, then not happen again for a couple of weeks).   And I have no convincing evidence or argument that it's 'Sesame', and not a hardward issue (network card, network cable, network router, security software, etc) ... except my notional vague observation that the problems increased after moving to 2.5.3 from 2.5.2 (or maybe it was going from 2.5 to 2.5.2).  I really don't have a clear enough chronology at this point to be very cogent about it - sorry!

I'll report in once I get some experiential history on the 'muscle fix'.

Hi Ray,

Will try to get that done this week.  Thanks!

Granted it sounds like a software problem, but re 6-year old computers, have you tried replacing the CMOS battery? Years ago I mentioned to my mother that the batteries last about 3 years. She said that months later the school district she worked for had a bunch of brand new computers and crazy things were happening. She happened to mention to the tech about the batteries and he said, "I completely forgot that! We actually bought these about 3 years ago and they've been sitting in storage.  They're not actually "new"." He replaced the batteries and the problems disappeared.

Also, after the battery, the next most common problem is a power supply that goes "flaky".  Usually the bearings will wear out and the fan gets noisy before that happens.  Years ago, twice, I figured, "I'm a tech! I can replace a fan!" HUGE mistake! Because it's mechanical, the fan is the weakest part. Once that's replaced, the second weakest part is the voltage regulator. It doesn't crater, it gets flaky. And depending on whether the computer is left on or not, 4-6 years is a common time frame.

Just a thought, not intended as a put-down...

Is all of your hardware used? I wonder if folks sell troublesome computers, like they do with cars that are lemons. If so, you may be getting a higher percentage of flaky hardware components than the industry average because you could be buying other peoples' problems?

My problem with the "old computer" theory is that Sesame is developed on and tested on several very old computers. Until recently the primary development box was a Pentium 3. One of the current development boxes is a ten year old Pentium 4. Even the newest development box is an older Xeon (5110). Here at Hammer Data / Lantica, we test across a broad range of boxes from Pentium 2 up through the i3/i5/i7 series.

Unlike Q&A, which ran right on top of the hardware, as was common for DOS programs, Sesame is built on top of cross-platform libraries. This allows us to be easily ported to several different operating systems. It also prevents Sesame from becoming reliant on any particular hardware / software configuration. We have to write very bland vanilla code to be able to run on both Windows and Linux using almost entirely the same exact code.

So, while it may be possible that older hardware is the problem, it isn't so likely. Unfortunately, Windows is susceptible to creeping damage. This means that as you use it, the OS becomes more and more broken. Device drivers gradually can corrupt. And, like any other program that uses a device, Sesame has to interface with the device driver for that device. We don't talk to it directly, but through an interface layer. Nor do we "timeout" on printer drivers. It may be possible that the OS is timing out and throwing an exception while we are trying to access a printer, especially one through the net. We have not seen this here, nor have we received reports indicating that this is common. We do timeout on our own network communications between client and server, if there are network problems. Older computers often have older network cards and these can deteriorate over time. There is some possibility that, with overheating, other components are having issues. Keep your vents clear.

As I mentioned early, it looks to me like you have at least one problem affecting your server and another one or two on the client. I believe the likeliest explanation for why it affects some computers more than others is that different computers run at differing speeds and thereby produce different timings. I recommend that you do "freshen" the OS on your older boxes (backup everything and re-install), just to clear up any creeping clutter that is impairing the OS. You should also check your logs for any network errors coming from old or bad network cards.

Another thing if you do an OS reinstall is try the software before adding all the updates.  Just do a basic XP Svc Pack 3 install, no updates, and try it.  Then download a copy of War and Peace, go get a large cup of coffe, and do all the updates. After you finish reading you should have time for a second coffee. 

We still use QA, running XP as the server and awhile back all the Win 7 workstations suddenly couldn't access QA on the server.  We found out that an XP update was incompatible.  Removed the update and everything works.