| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 806.1 | 46 hours is a very long time | TOOK::DITMARS | Pete | Mon Mar 18 1991 10:07 | 11 |
| PTB most definitely does not usually take that long.
Is there a lot of other activity on the system?
Could you do a show proc/all on the process that is running PTB?
Is PTB making progress, i.e. do you occasionally get new "Processing entity"
messages and the entity numbers seem to be increasing?
Are your dictionary/target parse table files being accessed locally, or via
the network?
|
| 806.2 | Is it pagefaulting like mad ? | WIKKIT::WARWICK | Trevor Warwick | Mon Mar 18 1991 12:42 | 9 |
|
I recently saw a PTB get almost completely stuck. It was on a
memory-tight timesharing system, and it ended up doing 10,000
pagefaults per second continually for about 6 hours. In this time, it
processed about 2 entities.
I eventually had to restart it with some more WSQUOTA.
Trevor
|
| 806.3 | WSQUOTA is too low? | SUBWAY::JENG | | Mon Mar 18 1991 15:34 | 13 |
| The PTB rebuild process is finally done with a successful message
(48 hours total). The TCP/IP AM installation IVP is also successful.
This is a 8M VAXstation 3100. The pagefaults value was over 25
millions, and the CPU time was 4:35:16. During PTB rebuild process,
the only background activity was the clock. Besides, the parse
tables are located locally. I think Trevor makes the point, the
WSQUOTA was too small (4000 blocks).
Could someone tell me how long does it take him/her to rebuild the
parse tables?
/Eng
|
| 806.4 | 4 hours on my system | HAWK2::GOLDMAN | Amy Goldman, DTN 226-5115 | Tue Mar 19 1991 09:19 | 5 |
| On my 16MB VAXstation 3100, the longest it took was around 4
hours, and that was with other processes going on. WSquo for
the account I installed from is only at 2048.
amy
|
| 806.5 | definitely sounds like memory
| TOOK::CALLANDER | | Tue Mar 19 1991 16:39 | 14 |
| I run a regular VAXstation/GPX with a full complement of memory, and using
the minimum required quotas I run in about 2 and 1/4 hours if nothing much
else is running; if nothing is running I have been known to run in about
2 hours even. The longest run I have had was about 5 hours and I had my
system pretty well loaded down.
Usually when PTB runs for more than a few hours you are having memory
problems; it is easier to stop and restart it then to let it finish.
I have seen it take over 24 clock hours to complete when there was not
enough memory to do it all; I have also seen if fail on the write to disk
because there wasn't enough memory for the last allocation.
|