Hello!
My problem aroused after installing a new 3-way raid controller and a new
10/100Mb network interface card into an alphaserver 400.
In order to connect tu0 to my net I had to upgrade fra DU v4.0B to v4.0D
at the same time.
The problem is: After serving NFS- and samba-services for dome time the
server reports;
Nov 18 13:43:58 harald vmunix: malloc_mem_alloc: no space in map (*)
and NFS and CIFS-services do no respond.
I'm able to log into the machine with ssh or telnet and make a reboot and
then the services is up once more.
No coredump, neither an entry in the uerf-log is to be found.
Only the line (*) from /var/adm/messages.
In order to fix the problem the machine is upgraded from 64MB to 192MB ram
without signifcant change.
I've tried to use immediate mode for swap space allocation without any
success. Swap space is 384MB without change. When I check swapon reports
3-5% in use
The 3-way raid controller is connected to a StorageWorks-box (a
350-something) with 7*RZ29B (narrow) in a raid5-diskvolume.
Regards,
Kjell Andresen Systems administrator, University of Oslo, Norway
Center for Information Technology Services and
Department of Geophysics
PS: I've searched many places for malloc_mem_alloc without a hit.
-----------------------------------------------------------------------
Here is the log:
Firstly: Dates of trouble, namely #1 to #14
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
harald /var/adm# grep malloc messages
Jul 16 16:16:28 harald vmunix: malloc_mem_alloc: no space in map #1
Jul 29 16:16:54 harald vmunix: malloc_mem_alloc: no space in map
Aug 7 10:45:03 harald vmunix: malloc_mem_alloc: no space in map
Aug 14 09:55:13 harald vmunix: malloc_mem_alloc: no space in map
Aug 21 11:40:24 harald vmunix: malloc_mem_alloc: no space in map #5
Sep 7 14:18:31 harald vmunix: malloc_mem_alloc: no space in map
Sep 17 14:54:47 harald vmunix: malloc_mem_alloc: no space in map
Sep 29 18:26:49 harald vmunix: malloc_mem_alloc: no space in map
Oct 13 18:30:08 harald vmunix: malloc_mem_alloc: no space in map
Oct 21 15:16:05 harald vmunix: malloc_mem_alloc: no space in map
Oct 30 10:32:06 harald vmunix: malloc_mem_alloc: no space in map #11
Nov 9 16:01:36 harald vmunix: malloc_mem_alloc: no space in map
Nov 11 11:31:42 harald vmunix: malloc_mem_alloc: no space in map #13
Nov 18 13:43:58 harald vmunix: malloc_mem_alloc: no space in map
Secondly: Actions and Problem solving:
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
#0 Wed Jun 17 16:34:51 1998: xcr0, the raid-set and the NIC
DE-500 set up. v4.0B, 64MB ram
swapmode is deferred mode.
Wed Jun 17 20:50:14 1998: Upgraded to DU v4.0D. tu0 up
(DEC TULIP (10/100))
#1 Thu Jul 16 16:49:54 1998: First time malloc_mem_alloc error
(July is holyday month here, (*))
#5 Fri Aug 21 12:09:30 1998: upgraded to 192MB ram
#11 Fri Oct 30 11:45:46 1998: one fileset on raid_domain 112% full
Tue Nov 3 07:26:15 1998: New swapmode: immediate mode set
#13 Mon Nov 9 17:16:01 1998: Back to deferred mode swapping.
(*) November is just before exam and delivery of main subjects..
Thirdly: Some swap statistics:
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
harald /var/adm# swapon -s
Swap partition /dev/rz0b:
Allocated space: 49152 pages (384MB)
In-use space: 1486 pages ( 3%)
Free space: 47666 pages ( 96%)
Total swap allocation:
Allocated space: 49152 pages (384MB)
In-use space: 1486 pages ( 3%)
Available space: 47666 pages ( 96%)
harald /var/adm# man pstat
No reference page found for pstat.
harald /var/adm# vmstat -P
Total Physical Memory = 192.00 M
= 24576 pages
Physical Memory Clusters:
start_pfn end_pfn type size_pages / size_bytes
0 249 pal 249 / 1.95M
249 24575 os 24326 / 190.05M
24575 24576 pal 1 / 8.00k
Physical Memory Use:
start_pfn end_pfn type size_pages / size_bytes
249 277 unixtable 28 / 224.00k
277 283 scavenge 6 / 48.00k
*
277 280 scavenge 3 / 24.00k
*
283 673 text 390 / 3.05M
673 750 data 77 / 616.00k
750 862 bss 112 / 896.00k
863 983 kdebug 120 / 960.00k
983 988 cfgmgmt 5 / 40.00k
*
990 1035 unixtable 45 / 360.00k
*
1048 1410 vmtables 362 / 2.83M
1410 24575 managed 23165 / 180.98M
============================
Total Physical Memory Use: 24313 / 189.95M
Managed Pages Break Down:
free pages = 974
active pages = 808
inactive pages = 777
wired pages = 5340
ubc pages = 15272
==================
Total = 23171
WIRED Pages Break Down:
vm wired pages = 1086
ubc wired pages = 1628
meta data pages = 1459
malloc pages = 786
contig pages = 11
user ptepages = 343
kernel ptepages = 19
free ptepages = 8
==================
Total = 5340
harald /var/adm# uptime
14:10 up 1 day, 17 mins, 3 users, load average: 0.01, 0.05, 0.05
Received on Thu Nov 19 1998 - 13:14:13 NZDT