Operating System - HP-UX 中的主題 Re: sar errors

sar errors

Matt Hearn — Fri, 07 Nov 2003 10:16:20 GMT

Hey folks; I'm trying to run "sar" on an 11.11 box, and I'm getting this response:

HP-UX ourhost B.11.11 U 9000/871 12/31/69

%usr %sys %wio %idle
40 20 20 20
2 2 95 2
sar: realloc() error

Anybody know what this means, and how to fix it? Thanks!!!

Re: sar errors

Mark Grant — Fri, 07 Nov 2003 10:20:59 GMT

sar is having trouble obtaining enough RAM to do the job. It could be you are running out of memory or it might be a kernel parameter issue.

Or, I could be wrong :)

Re: sar errors

Geoff Wild — Fri, 07 Nov 2003 10:26:07 GMT

I have to agree with Mark - sounds like you are out of resources...

You may only need to increase the kernel parameter: maxssiz

Rgds...Geoff

Re: sar errors

Michael Steele_2 — Fri, 07 Nov 2003 11:23:05 GMT

'realloc' expands the memory buffer or resizes with a read operation which should be pretty automatic from the O/S.

For you to suddenly see this error may indicate another problem, i.e., lost some memory.

What's showing up in syslog.log, EMS and STM?

How's your swap?

How's your 'sar' patches?

s700_800 11.11 sar(1m) patch PHCO_24477
notes PHCO_24477
notes
s700_800 11.11 sar shows incorrect values for avwait, avque PHKL_27200
notes PHKL_27200
notes
s700_800 11.11 SCSI IO Cumulative Patch PHKL_29047
notes PHKL_29047
notes
s700_800 11.11 VxFS cumulative patch, perf of mmap()d files PHKL_27830
notes PHKL_29110
notes
s700_800 11.X ANSI C compiler B.11.11.04 cumulative patch PHSS_26792 PHSS_26792
s700_800 11.X +O4/PBO Compiler B.11.11.04 cumulative patch PHSS_26793 PHSS_26793

Re: sar errors

James Randall — Fri, 07 Nov 2003 14:55:18 GMT

I would run "swapinfo".

The line that is important is:

memory ###### ######## ##### %%%

Look at the percentage line....If this is close to 100.....Then you probably need more RAM.

Re: sar errors

Matt Hearn — Fri, 07 Nov 2003 15:03:25 GMT

Nah, memory is okay:
memory 1204112 418368 785744 35%

I'll try and schedule a time to apply those patches, though.

Re: sar errors

Geoff Wild — Fri, 07 Nov 2003 15:05:10 GMT

Run this command - just discovered yesterday:

/usr/contrib/Q4/bin/kmeminfo

Output like:

kmeminfo (3.11)
libp4 (5.17): Opening /stand/vmunix /dev/kmem

Loading symbols from /stand/vmunix

======================================================================
Date: Fri Nov 7 13:03:18 2003

Processing pfdat table (2505108 entries)...

----------------------------------------------------------------------
Physical memory usage summary (in pages):

Physmem = 2580480 Available physical memory:
Freemem = 799983 Free physical memory
Used = 1780497 Used physical memory:
System = 333474 by kernel:
Static = 75372 for text and static data
Dynamic = 0 for dynamic data
Bufcache = 258048 for file-system buffer cache
Eqmem = 54 for equiv. mapped page pool
SCmem = 0 for system critical page pool
User = 959337 by user processes
Uarea = 8112 for thread uareas
Disowned = 63983 disowned pages

----------------------------------------------------------------------
Dynamic memory usage summary (in pages):

Dynamic = 0 Kernel dynamic data (sysmap):
MALLOC = 0 memory buckets:
bucket[ 5] = 0 size 32 bytes
bucket[ 6] = 0 size 64 bytes
bucket[ 7] = 0 size 128 bytes
bucket[ 8] = 0 size 256 bytes
bucket[ 9] = 0 size 512 bytes
bucket[10] = 0 size 1024 bytes
bucket[11] = 0 size 2048 bytes
bucket[12] = 0 size 4096 bytes
bucket[13] = 0 size 2 pages
bucket[14] = 0 size 3 pages
bucket[15] = 0 size 4 pages
bucket[16] = 0 size 5 pages
bucket[17] = 0 size 6 pages
bucket[18] = 0 size 7 pages
bucket[19] = 0 size 8 pages
bucket[20] = 0 size > 8 pages
Kalloc = 0 kalloc()
Eqalloc = 0 eqalloc()
Reserved = 0 Reserved pools

Re: sar errors

Matt Hearn — Wed, 19 Nov 2003 15:27:35 GMT

Okay, I installed PHCO_24477 on the three boxes. On one machine (D370) it worked great, thanks!

But on the other two, a cluster of D390s, I now get a coredump from sar instead of the realloc() error. Memory is steady at about 39% on both machines.

Any more ideas?

Re: sar errors

Michael Steele_2 — Wed, 19 Nov 2003 23:20:10 GMT

Let's do a couple of things. Can you attach:

a) file core
b) strings core | more

And the errors from :

c) swverify -x=check_permissions=ture \*
d) check_patches

For c & d refer to the swverify.log file.