Yesterday we had 2 batch processes that crashed with an error I've never seen before. The protrace files seem incomplete to me . It happened on 2 seperate servers connected to separate databases.
OE 11.3.2, Windows 2012 R2
It was for the same customer using servers in the same datacenter.
For a something that happened only once I don't want to start a tech support case, but I'm still curious if someone has seen this type of error before and what could cause it.
The only link I found with google http://www.peg.com/forums/dba/200408/msg00423.html
1st file , not trimmed
13/07/15 21:07:09 [7156] Progress Recent Message(s): (7129) (452) Usr 31 set name to batchclnt. (7129) Login by TDU_LISA_PRO on CON:. (452) *** ABL Call Stack *** Last action: TY_IOPRT (81) 8198: Output-Export-Bestand xmencm/agingex-p.p (.\gui\xmen\xmencm\agingex-p.r) 7838: Export-Aging-Detail xmencm/agingex-p.p (.\gui\xmen\xmencm\agingex-p.r) 7528: Export-Aging xmencm/agingex-p.p (.\gui\xmen\xmencm\agingex-p.r) 7361: xmencm/agingex-p.p (.\gui\xmen\xmencm\agingex-p.r) 1945: ccetools/start-b.p (.\gui\ccetools\start-b.r)
2nd file, not trimmed
13/07/15 20:32:37 [4964] Progress Recent Message(s): *** ABL Call Stack *** Last action: TY_IOPRT (81) 9906: OUTPUT-ANALDB-DWH xmen/menana-p.p (.\gui\xmen\xmen\menana-p.r) 10463: OUTPUT-CSV-BESTAND-DWH xmen/menana-p.p (.\gui\xmen\xmen\menana-p.r) 6991: xmen/menana-p.p (.\gui\xmen\xmen\menana-p.r) 1939: ccetools/start-b.p (.\gui\ccetools\start-b.r)
To be honest I'd be logging a support case, because, as you say, there is just that sole hit on PEG about the problem.
Has anything changed in your environment recently? OS? Windows Updates?
Those look like procore files, not protrace.
The "Last action: TY_IOPRT (81)" refers to the internal action codes in the compiled r-code.
Call stack should still match up with debug listing though, so check what's on xmencm/agingex-p.p, line 8198 and xmen/menana-p.p, line 9906 and that should tell you which ABL statement triggered this.
Definitely protrace files and not procore. See if you can locate the actual procore files. Also make sure to look in the database log, and the windows event logs.
I would assume TY_IOPRT has something to do with writing to disk. So odds are good that you are going to find something related to streams when you look at those lines of code.
Are these users running on a terminal server or local to their PC?
> I would assume TY_IOPRT has something to do with writing to disk
The action codes with nearest numbers:
79. TY_BRE 80. IN/OUT CLOSE 81. TY_IOPRT 82. CHOOSE 83. UNDERLINE 84. COLOR 85. TY_SCHMASG 86. ASSIGN
Indeed it's very likely that TY_IOPRT is related to I/O.
Also it's a very old action code (V10.2B has 149 action codes).
Typically Tech support will ask if I can reproduce the issue, which I can't. If it occurs more often I surely will report it.
In this case I suspect some issue in the datacenter as it occured on 2 different servers.
A common factor is that both procedures were exporting data to CSV files.
Thanks for all the replies, I'll post further updates if I find more information about it.