Fellow Listers,

We have a process that daily selects any spoolfiles sent to a specific device
class, we compress them with LZW and transfer them to an outside vendor to
create microfiche slides.  First the LZW 'archive' is sent in binary mode and
then a second file, that represents the 'directory' of the first file, in ASCII.

On a number of occasions, specifically when the 'archive' is rather large, the
transfer of the 'archive' finishes (this is determined by monitoring the record
being accessed using GLANCE) but the ftp process does not continue with the
transfer of the 'directory' file.  Instead it 'sits there' until some time out
is detected and the rest of the job is aborted.

I have taken steps to minimize these occurrences by incorporating the size of
each spoolfile, in terms of pages, and group in separate 'archives' spoolfiles
within specific size.

Well last night the process hung again because there were too many spool files
in the '>=5000' page range, and today I executed the submitted command file
manually.  The process got hung just after the 'current record' reached the
file's eof (from Glance), the data file is not being accessed any more but the
ftp process has not continued with the transfer of the 'directory' file.  The
'showproc' command also shows that the ftp process is NOT accumulating any more
'cputime' while Glance shows the process in a wait state of 'other' and the
following stack trace:



Procedure Trace for Pin  210 is:

       PC=a.0016870c enable_int+$2c
NM* 0) SP=41847970 RP=a.01279ec0 notify_dispatcher.block_current_process+$324
NM  1) SP=41847970 RP=a.0127c348 notify_dispatcher+$264
NM  2) SP=418478f0 RP=a.0037b418 ipc_impede+$274
NM  3) SP=418477f0 RP=a.0037b190 ?ipc_impede+$8
         export stub: a.01f06300 sk_block+$1c4
NM  4) SP=418476b0 RP=a.01f06c40 sk_block_for_completion+$fc
NM  5) SP=41847570 RP=a.01f41df4 sk_recv+$25c
NM  6) SP=41847430 RP=a.01eea600 IPCRECV+$2fc
NM  7) SP=41847370 RP=a.01eea2d0 ?IPCRECV+$8
         export stub: e31.00022cbc
NM  8) SP=41847130 RP=e31.000231ac
NM  9) SP=418470b0 RP=e31.0002b748
NM  a) SP=41846df0 RP=e31.0002d430
NM  b) SP=41846b30 RP=e31.00034fac
NM  c) SP=418461b0 RP=e31.00000000
     (end of NM stack)


Press RETURN to continue (or enter a GLANCE command):

Procedure Trace for Pin  210 is:

       PC=a.0016870c enable_int+$2c
NM* 0) SP=41847970 RP=a.01279ec0 notify_dispatcher.block_current_process+$324
NM  1) SP=41847970 RP=a.0127c348 notify_dispatcher+$264
NM  2) SP=418478f0 RP=a.0037b418 ipc_impede+$274
NM  3) SP=418477f0 RP=a.0037b190 ?ipc_impede+$8
         export stub: a.01f06300 sk_block+$1c4
NM  4) SP=418476b0 RP=a.01f06c40 sk_block_for_completion+$fc
NM  5) SP=41847570 RP=a.01f41df4 sk_recv+$25c
NM  6) SP=41847430 RP=a.01eea600 IPCRECV+$2fc
NM  7) SP=41847370 RP=a.01eea2d0 ?IPCRECV+$8
         export stub: e31.00022cbc
NM  8) SP=41847130 RP=e31.000231ac
NM  9) SP=418470b0 RP=e31.0002b748
NM  a) SP=41846df0 RP=e31.0002d430
NM  b) SP=41846b30 RP=e31.00034fac
NM  c) SP=418461b0 RP=e31.00000000
     (end of NM stack)


Press RETURN to continue (or enter a GLANCE command):

Procedure Trace for Pin  210 is:

       PC=a.0016870c enable_int+$2c
NM* 0) SP=41847970 RP=a.01279ec0 notify_dispatcher.block_current_process+$324
NM  1) SP=41847970 RP=a.0127c348 notify_dispatcher+$264
NM  2) SP=418478f0 RP=a.0037b418 ipc_impede+$274
NM  3) SP=418477f0 RP=a.0037b190 ?ipc_impede+$8
         export stub: a.01f06300 sk_block+$1c4
NM  4) SP=418476b0 RP=a.01f06c40 sk_block_for_completion+$fc
NM  5) SP=41847570 RP=a.01f41df4 sk_recv+$25c
NM  6) SP=41847430 RP=a.01eea600 IPCRECV+$2fc
NM  7) SP=41847370 RP=a.01eea2d0 ?IPCRECV+$8
         export stub: e31.00022cbc
NM  8) SP=41847130 RP=e31.000231ac
NM  9) SP=418470b0 RP=e31.0002b748
NM  a) SP=41846df0 RP=e31.0002d430
NM  b) SP=41846b30 RP=e31.00034fac
NM  c) SP=418461b0 RP=e31.00000000
     (end of NM stack)


Press RETURN to continue (or enter a GLANCE command):


If I was to let the process go, I know that eventually I'll get the following
FTP related errors:

FTPLASTERR = 66
FTPLASTREPLY = 150 Opening BINARY mode data connection for hts_991008D.dat.
FTPLASTMSG = INVALID CONNECTION DESCRIPTOR.  (SOCKERR 66)

Any Ideas?

Regards
Paul H. Christidis