[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference csc32::consolemanager

Title:POLYCENTER Console Manager
Notice:Kits, Scans, Docs on CSC32:: as PCM$KITS:,PCM$DOCS:, PCM$SCANS:
Moderator:CSC32::BUTTERWORTH
Created:Thu Aug 06 1992
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:1541
Total number of notes:6564

1180.0. "Still controler LEF" by 51564::WESTERBACK (Mai pen rai!) Thu Feb 01 1996 04:03

    We still have the same problem described several times before (e.g.
    in note 1070).
    
    After PCM has been running for a couple of hours one or two of the
    contoller processes goes into LEF. Can't do monitor or connect to
    the systems on that controller.
    
    PCM 1.6-119 on an Alpha 3000-600 VMS 6.1 with 128 Mb memory.
    41 systems connected, via Decserver 200.
    Used for internal IS systems in Sweden.
    
    I have also installed the new images (dated 18-dec-1995) announced 
    in this conference as ALPH61.BCK.
    
    Have raised CHANNELCNT to 1000 as recommended, didn't help.
    Archiving which was done via DECscheduler has been stopped since
    a few days. There is plenty of disk free, so no auto-archiving
    ought to be done, I guess. But we still se shortlived "Archive"
    processes once in a while. Especially when we shutdown PCM and
    restart there are sometimes a lot of them. Any idea why?
    
    Since this Alpha boots VERY fast, it's a quick fix to reboot, but it
    seems that earlier we had a hang every one or two days, but now it
    seems to only work a few hours before hanging.
    
    Is there anything more we can try to get this working? Will the awaited
    ECO-2 solve this kind of problem?
    
    Rgds,
    Hans
T.RTitleUserPersonal
Name		DateLines
1180.1More details51564::WESTERBACKMai pen rai!Thu Feb 01 1996 09:5326
    On closer examination this might be two different problems:
    
    At first it stopped working about once a day. Could not do a monitor
    or connect. I wasn't aware of the controller LEF problem at that time.
    But still you could do a CONS EXTRACT and get the data up to current
    time. Also after reboot, you could scroll back in cons monitor and see
    all data, except the short reboot interval. Some of the events came in
    a jumbled order, not chronological. So the data was kept by PCM
    but couldn't be presented correctly.
    
    Now it only works maybe two hours, and extract doesn't work either.
    After restart of PCM or reboot, there is no data kept for the recent
    hours, while monitor/connect wasn't working.
    
    I'm not sure what has changed. The upgrades were done earlier, so
    that's no difference. We have stopped archiving that's all, and fiddled
    around a bit with CHANNELCNT and TTY parameters.
    
    When it was once a day we thought we could get around it with a daily
    reboot done by DECscheduler, but now when it gets down to only a few
    hours PCM is getting quite unusable. 
    
    Any input from others who have gotten around these problems?
    
    Rgds,
    Hans
1180.2CSC32::BUTTERWORTHGun Control is a steady hand.Thu Feb 01 1996 11:3510
    Your running a rather old build. We have fixed several hang situations
    so please get the following saveset:
    
    CSC32::DUMPS:[BUTTERWORTH.PCM]CSC_126_AXP.BCK
    
    There is also a RESTORE_COMMANDS_FOR_BCKS.TXT file that gives you the
    necessary BACKUP commands to restore the saveset.
    
    Regards,
       Dan
1180.3OK51564::WESTERBACKMai pen rai!Fri Feb 02 1996 09:4811
    OK, we've been up and running for 6 hours, looks good so far.
    
    Still a question: Seen archiving processes during the day. At one time
    there were about 10 of them, running up 100% CPU. Why do these appear?
    Disks are nowhere near 80 % full.
    
    Is there still an archiving problem, or can we turn on regular
    archiving?
    
    Rgds,
    Hans
1180.451564::WESTERBACKMai pen rai!Thu Feb 08 1996 04:2017
    OK, our PCM seemed to be working OK for a few days. As I said we had
    disabled archiving.
    
    This morning I noticed that one of the three Console Ctrl process had
    disappeared. Accounting showed an acc vio at 08:04. I also noticed
    archiving processes which had started at 07:30 (and kept going until
    10:00). It seems that these archiving processes were started because
    we hit 80% on the log disk.
    
    After the archiving was done, I restarted PCM with console$startup
    restart, and get all processes back. BUT can the acc vio have something
    to do with the archiving ?  I will put the accounting info in next
    reply.
    
    Hans
1180.5accounting51564::WESTERBACKMai pen rai!Thu Feb 08 1996 04:2022
    
DETACHED Process Termination
----------------------------

Username:          SYSTEM            UIC:               [SYSTEM]
Account:           SYSTEM            Finish time:        8-FEB-1996 08:05:56.07
Process ID:        000000B6          Start time:         7-FEB-1996 02:51:47.08
Owner ID:                            Elapsed time:                1 05:14:08.99
Terminal name:                       Processor time:              0 01:42:37.20
Remote node addr:                    Priority:          4
Remote node name:                    Privilege <31-00>: FFFFFFFF
Remote ID:                           Privilege <63-32>: FFFFFFFF
Queue entry:                         Final status code: 1000000C
Queue name:
Job name:
Final status text: %SYSTEM-F-ACCVIO, access violation, reason mask=!XB, virtual

Page faults:          3045996        Direct IO:            1751623
Page fault reads:          30        Buffered IO:          2674902
Peak working set:        5344        Volumes mounted:            0
Peak page file:         20512        Images executed:            3