Pernicious Winsock Errors (Empower)

[DR]

Anyone having frequent errors where an eSATIN just quits collecting data for no apparrent reason? I've got 2 eSATINs plugged into 1 LAC/E via a MiLAN switch & instrument LAN connection on a LAC/E.
I've tried 2 different LAC/Es, different switches, different cables & even tried a different port for the LAC/E-LAN connection.

other suggestions?

[bartjoosen]

some ideas:
did you disabled windows error reporting and drwtsn32?
did you try to delete your preferences (sometimes this really works)?
did you setup a fixed page file size of at least 500mb?

Bart

[DR]

Yes

Not sure what you mean by deleting preferences

I found that witht he current LAC/E, the swapfile had mysteriously been set to 0 (I locked it in at 4096), the previous LAC/E in that spot had a decent swap file...

[bartjoosen]

By deleting preferences, I mean: you can save your preferences like in quickset, review, .... there is also a menu item "delete preferences", sometimes it helps by clicking this menu item. This helped us in some occasions where the software quits for some reason. It has something to do with a table of preferences which is full. If you flush this table, it should be ok.

Bart

[DR]

I tried it, but I have my doubts as to how helpful it will be. The LAC/E is new and we run in a client/server (Citrix) configuration, so nobody has a chance to save any preferences on that LAC/E's copy of Empower...

[bartjoosen]

It was just worth trying, sometimes it helps, most of the time not

I'm not familliar with citrix based configurations, but does the Empower gives an error in the message center, or on screen?

Bart

[DR]

Errors show in msg ctr. and the person moitoring the run from his desktop citrix client sees "instrument error" on his status line & "Start/Stop Run" button goes from red to green on that SS.

[bartjoosen]

What error is in the msg cntr? Or just the really informative "instrument failure"?

Your MiLAN switch isn't warming up?

Bart

[DR]

I switched switches...

typical message set:
3761 Warning Instrument
eSATIN#L04...: (2)A failure occured in reading the reply from Get Data.
3760 Warning Instrument
eSATIN#L04...: (1)The socket failed while reading data with error code: WinSock Error # 10055
3759 Error Instrument
Instrument Failure eSATIN#L04...

[bartjoosen]

lol, I'm out of ideas, maybe you should call Waters?

Good luck

Bart

[DR]

I have called Waters.
So far, my best guess is that there's something on the LAN here that eSATINs are sensitive to that the older SAT/INs aren't sensitive to as I've found no appreciable differences in LAC/E configurations between reliable and unreliable combinations of LAC/Es & eSATINs.

[HbJ]

That error can have several reasons (speaking as a computer scientist):


- Which version of Windows are you using?
- How many ohter applications are running?
- What is your available disc space? (As this may be too low)
- How many resources do you have free when the error occurs?
- How much RAM does your box have?


Why am I asking these question?
The error message you are seeing (Winsock error #10055) points towards low system resources (GDI) and/or insufficient memory and/or a lack of sockets.

Maybe the older version of your software used a slightly different frame format which is more efficient or it just transmitted less data?

[bartjoosen]

HbJ is right, WinSock Error # 10055 is due to low resources.

Maybe increase paging file size (look after available disk space) to 1500 mb?

[DR]

- Which version of Windows are you using?
XP w/ SP2
- How many ohter applications are running?
None - it's a dedicated LAC/E (the later of 2 current models)
- What is your available disc space? (As this may be too low)
~75GB, 87% free, fragmentation very low, page file is fixed at 4080MB
- How many resources do you have free when the error occurs?
Don't know
- How much RAM does your box have?
1GB
The error message you are seeing (Winsock error #10055) points towards low system resources (GDI) and/or insufficient memory and/or a lack of sockets.
I thought it was because the instrument NIC lost track of a device.

Maybe the older version of your software used a slightly different frame format which is more efficient or it just transmitted less data?
We've only had SPs G-J added since original installation.

[HbJ]

Oh heck...

As I'm not familiar with the software you're using I'd be very glad if you could fill me in about your configuration, especially how the network structure looks like and how the computers get their IP addresses.

Another possibility could be too long network segments which lead to late collisions. Do you use 100MBit or Gigabit Ethernet?

The first error message
"(2)A failure occured in reading the reply from Get Data. "
make me think. Who throws this error?

It seems like a paket gets mangled, your software can't cope with that error and runs out of data (?).

How do your network nodes agree about the MTU?

Are there any special conditions when the error occurs (start of run, middle of run, peak, flat baseline, ...)?

How long does it take from the first to the second error?

Please run "netstat -an >protocol.txt" (without the quotes) the next the error occurs so I can see if you're out of sockets.

And try to get the hotfix for http://support.microsoft.com/kb/905628/en

Maybe a station is losing its connection and tries to resend all data at once. That huge TCP/IP stream leads to an exhaustion of the receivers resources.

Do you get _any_ feedback from Waters regarding that problem?

And when did it occur the first time?
Did you connect new devices to the network?
And what file system are you using?

I know this looks like 'Dr. House: In search of a solution' but this error seems strange to me :\