Re: Witango-Talk: Horrendous WT5 problems
Roland Dumas at [EMAIL PROTECTED] wrote: > that's what the ODBC-related crashes look like. Might be related to > when the datasource connection expires. Hmmm, less than what I would call robust. Is there a solution to the ODBC problem? That said, there have been some events since. I eventually decided to restart the server, but it had other ideas. It got as far as loading the menu bar and then got into a cycle of flashing the menu bar on and off repeatedly. I tried starting in Safe mode but didn't know enough about the shell to do anything useful. Tried several other key holds but to no avail. I figured I could do a repair with the system disks but no, it wants to erase the volume. Eventually I started in Target disk mode (yes, an X Serve will do that) and daisy chained my Powerbook - X Serve - 30Gb drive and did a back up to the drive. I then reinstalled OS X Server, WebStar, Witango, FileMaker and all the other bits. Did much of it with the installers (from the back up drive), but also copied over the 'Webserver' folder and various Witango License files and configs. The current set up hasn't crashed yet. The Witango process shows as Hung in the activity monitor but still runs. Anybody know what a 'Hung' process actually is. This has been a fairly painful process and 10 months of (part time) work and many thousands of aussie dollars spent and I'm still not convinced it's stable. Thank you for your feedback Roland. It would have been a fairly quiet and lonely process otherwise (hint to Witango support ;) ) Wayne Irvine Byte Services Pty Ltd http://www.byteserve.com.au/ [EMAIL PROTECTED] Ph 02 9960 6099 Mob 0409 960 609 Fax 02 9960 6088 TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf
Re: Witango-Talk: Horrendous WT5 problems
that's what the ODBC-related crashes look like. Might be related to when the datasource connection expires. On Jul 28, 2004, at 10:40 PM, Wayne Irvine wrote: I am beginning to suspect it's not a TAF or traffic problem. The problem seems to happen when I'm not testing it and at this stage I'm the only one hitting it (presumably). I saw that it was running before I went to lunch, but when I got back it was no longer running. Wayne Irvine Byte Services Pty Ltd http://www.byteserve.com.au/ [EMAIL PROTECTED] Ph 02 9960 6099 Mob 0409 960 609 Fax 02 9960 6088 ___ _ TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf
Re: Witango-Talk: Horrendous WT5 problems
I am beginning to suspect it's not a TAF or traffic problem. The problem seems to happen when I'm not testing it and at this stage I'm the only one hitting it (presumably). I saw that it was running before I went to lunch, but when I got back it was no longer running. Wayne Irvine Byte Services Pty Ltd http://www.byteserve.com.au/ [EMAIL PROTECTED] Ph 02 9960 6099 Mob 0409 960 609 Fax 02 9960 6088 TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf
Re: Witango-Talk: Horrendous WT5 problems
one more suggestion: look carefully at the results of your global search and replace. I had one case where an additional character occurred that tripped crashes. On Jul 28, 2004, at 3:45 PM, Wayne Irvine wrote: Roland Dumas at [EMAIL PROTECTED] wrote: You haven't converted all your tafs to get rid of the evil ¥ (yen) All the TAFs under T2000 talk to remote hosts so the paths were very different. I did multifile search and replace to change the datasources to #. making sure to save in Unix ISO Latin 1 character set and not Mac character set. Hmmm, this was something I was unaware of. Best check it out. Thanks Wayne Byte Services Pty Ltd http://www.byteserve.com.au/ [EMAIL PROTECTED] Ph 02 9960 6099 Mob 0409 960 609 Fax 02 9960 6088 ___ _ TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf
Re: Witango-Talk: Horrendous WT5 problems
On Jul 28, 2004, at 3:45 PM, Wayne Irvine wrote: Roland Dumas at [EMAIL PROTECTED] wrote: You haven't converted all your tafs to get rid of the evil ¥ (yen) All the TAFs under T2000 talk to remote hosts so the paths were very different. I did multifile search and replace to change the datasources to #. I did too, originally, and missed a couple. making sure to save in Unix ISO Latin 1 character set and not Mac character set. Hmmm, this was something I was unaware of. Best check it out. Thanks Wayne TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf
Re: Witango-Talk: Horrendous WT5 problems
Roland Dumas at [EMAIL PROTECTED] wrote: > You haven't converted all your tafs to get rid of the evil ¥ (yen) All the TAFs under T2000 talk to remote hosts so the paths were very different. I did multifile search and replace to change the datasources to #. > making sure to save in Unix ISO Latin 1 character set and > not Mac character set. Hmmm, this was something I was unaware of. Best check it out. Thanks Wayne Byte Services Pty Ltd http://www.byteserve.com.au/ [EMAIL PROTECTED] Ph 02 9960 6099 Mob 0409 960 609 Fax 02 9960 6088 TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf
Re: Witango-Talk: Horrendous WT5 problems
hypothesis 1: You haven't converted all your tafs to get rid of the evil ¥ (yen) sign. Tafs developed in prior versions of the studio (not sure which version changed) denote the "any source" for FMP with that character. They need to be changed to the # character, which won't kill witango. To do that, you need to open tafs in BBEdit or something like that and do the edits, making sure to save in Unix ISO Latin 1 character set and not Mac character set. If you've done that, you may have missed one. find which it is either by searching each taf in BBEdit for ¥ or by going through all your tafs to see which db access action crashes witango. (also run W* and W*mail on both machines that I run witango on and they've not been implicated in crashes ever. ) On Jul 28, 2004, at 2:56 PM, Wayne Irvine wrote: Roland Dumas at [EMAIL PROTECTED] wrote: Doesn't look like the crash logs I've seen, which are always associated with ODBC. are your data sources Filemaker? Yes. FIleMaker 5.5 How much RAM in the machine? 1Gb. It's also running 4D Webstar and 4D WebMail. Wayne Byte Services Pty Ltd http://www.byteserve.com.au/ [EMAIL PROTECTED] Ph 02 9960 6099 Mob 0409 960 609 Fax 02 9960 6088 ___ _ TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf
Re: Witango-Talk: Horrendous WT5 problems
Roland Dumas at [EMAIL PROTECTED] wrote: > Doesn't look like the crash logs I've seen, which are always associated > with ODBC. > are your data sources Filemaker? Yes. FIleMaker 5.5 > How much RAM in the machine? 1Gb. It's also running 4D Webstar and 4D WebMail. Wayne Byte Services Pty Ltd http://www.byteserve.com.au/ [EMAIL PROTECTED] Ph 02 9960 6099 Mob 0409 960 609 Fax 02 9960 6088 TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf
Re: Witango-Talk: Horrendous WT5 problems
Doesn't look like the crash logs I've seen, which are always associated with ODBC. are your data sources Filemaker? How much RAM in the machine? On Jul 27, 2004, at 10:50 PM, Wayne Irvine wrote: If anyone needs some bedtime reading following is the most recent crash log: TO UNSUBSCRIBE: Go to http://www.witango.com/developer/maillist.taf