Re: Counter overflow and memory problems

     I set these up on all my adsm servers to ease such number problems...
     We have servers that stay up 6 and more months (which I hate 'cause a
     periodic boot cures a lot of things) and I haven't seen that one...

                 Schedule Name: RESETBUFP
                   Description: Reset the Buffer Pool Statistics for the
                                 Database
                       Command: reset bufpool
                      Priority: 5
               Start Date/Time: 05/08/1998 00:01:00
                      Duration: 1 Hour(s)
                        Period: 1 Week(s)

                 Schedule Name: RESETDBMAXU
                   Description: Reset the Utilization Statistic for the
                                 Database
                       Command: reset dbmaxu
                      Priority: 5
               Start Date/Time: 05/08/1998 00:02:00
                      Duration: 1 Hour(s)
                        Period: 1 Week(s)

                 Schedule Name: RESETLOGCONS
                   Description: Reset the Cumulative Log Consumption Statistic
                       Command: reset logcons
                      Priority: 5
               Start Date/Time: 05/08/1998 00:03:00
                      Duration: 1 Hour(s)
                        Period: 1 Week(s)

                 Schedule Name: RESETLOGMAXU
                   Description: Reset the Utilization Statistic for the Log
                       Command: reset logmaxu
                      Priority: 5
               Start Date/Time: 05/08/1998 00:04:00
                      Duration: 1 Hour(s)
                        Period: 1 Week(s)

______________________________ Reply Separator _________________________________
Subject: Counter overflow and memory problems
Author:  tkssteve (tkssteve AT REXX.ACSU.BUFFALO DOT EDU) at unix,mime
Date:    8/4/98 4:07 PM


Hi All,

     My server has been running for 28 days now, and the q db f=d shows
that the counter for the buffer pool requests may have overflown (is that
a real word?).  It may be a coincidence, but last night, I had about a
dozen client session fail to connect due to "out of space" error messages.

Here is the DB info:

adsm> q db f=d

          Available Space (MB): 30,148
        Assigned Capacity (MB): 30,148
        Maximum Extension (MB): 0
        Maximum Reduction (MB): 2,740
             Page Size (bytes): 4,096
            Total Usable Pages: 7,717,888
                    Used Pages: 7,012,651
                      Pct Util: 90.9
                 Max. Pct Util: 90.9
              Physical Volumes: 32
             Buffer Pool Pages: 115,200
         Total Buffer Requests: -1,921,574,945
                Cache Hit Pct.: 0.00
               Cache Wait Pct.: 0.00
           Backup in Progress?: No
    Type of Backup In Progress:
  Incrementals Since Last Full: 2
Changed Since Last Backup (MB): 10.81
            Percentage Changed: 0.04
Last Complete Backup Date/Time: 08/04/98   16:00:02

The last number I remember seeing for buffer requests before it went
negative was around 2.1 billion.

And, here are some of last night's error messages:

ANR8210E Unable to establish TCP/IP session with 128.205.7.41 - connection
request timed out.
ANR9999D tcpcomm.c(375): Unknown TCPIP Failure. Return Code=148
_alloc_chunk(): _mmap failed: Not enough space
ANR0403I Session 24540 ended for node TERMINUS (SUN SOLARIS).
_alloc_chunk(): _mmap failed: Not enough space
_alloc_chunk(): _mmap failed: Not enough space
ANR8210E Unable to establish TCP/IP session with 128.205.7.41 - connection
request timed out.
_alloc_chunk(): _mmap failed: Not enough space
_alloc_chunk(): _mmap failed: Not enough space
ANR9999D pkthread.c(563): Thread creation failed; rc=12.
ANR9999D pvrmp.c(1752): Unable to start background dismount for volume
"00E9A5" - trying synchronous dismount.
ANR8336I Verifying label of 3570 volume 00E9A5 in drive D3 (/dev/rmt/4st).
ANR0403I Session 24539 ended for node ADCP1.CIT.BUFFALO.EDU (SunOS).
_alloc_chunk(): _mmap failed: Not enough space
_alloc_chunk(): _mmap failed: Not enough space
ANR0406I Session 24591 started for node ADCP1.CIT.BUFFALO.EDU (SunOS)
(Tcp/Ip
128.205.251.2(44029)).
ANR0403I Session 24591 ended for node ADCP1.CIT.BUFFALO.EDU (SunOS).
_alloc_chunk(): _mmap failed: Not enough space
ANR8468I 3570 volume 00E9A5 dismounted from drive D3 (/dev/rmt/4st) in
library
3575.
_alloc_chunk(): _mmap failed: Not enough space
_alloc_chunk(): _mmap failed: Not enough space
_alloc_chunk(): _mmap failed: Not enough space
ANR0403I Session 24552 ended for node LOG.CIT.BUFFALO.EDU (SunOS).
_alloc_chunk(): _mmap failed: Not enough space
_alloc_chunk(): _mmap failed: Not enough space

I am going to reboot the machine shortly.  Has anyone else seen these
types of error messages before?  Does ADSM have memory leaks?

Thanks,



Steve (unVMix Systems Programmer/Dude) Roder
(tkssteve AT ubvm.cc.buffalo DOT edu | tkssteve AT acsu.buffalo DOT edu | 
(716)645-3564 ,
   | http://ubvm.cc.buffalo.edu/~tkssteve)