HPE 9000 and HPE e3000 Servers
1748058 Members
5006 Online
108758 Solutions
New Discussion юеВ

Re: D390 memory error; change mem?

 
SOLVED
Go to solution
nkkorhaz
Advisor

D390 memory error; change mem?

hi,

this error messages from cstm-log:
-----------------------------------LOG:
Memory Error Log Summary
Error
Board Error Address Error Type Page Count
------------ ------------------ ---------- --------- -----
EXT0 0b 0x00000000284bfdf0 Single-Bit 0x00284bf 48959
EXT0 2a/2b 0x000000000989a000 Single-Bit 0x000989a 0
EXT0 0a/0b 0x000000003188c000 Single-Bit 0x003188c 0
EXT0 0a/0b 0x000000002f15e000 Single-Bit 0x002f15e 0

System start: Thu May 13 07:37:29 2004.
Last error check: Mon Sep 6 08:02:12 2004.
Logging interval: 3600 seconds.
4 address(es) with errors logged by memory logging daemon.

The Logtool Utility provides full details about the memory error log.

Page Deallocation Table (PDT)

Board Error Address Error Type Page
------------ ------------------ ---------- ---------
EXT0 0a/0b 0x000000002f15e000 Single-Bit 0x002f15e
EXT0 0a/0b 0x000000003188c000 Single-Bit 0x003188c
EXT0 2a/2b 0x000000000989a000 Single-Bit 0x000989a
EXT0 0a/0b 0x00000000284bf000 Single-Bit 0x00284bf

PDT Entries Used: 4
PDT Entries Free: 46
PDT Total Size: 50
-------------------------------END_OF_LOG

what can i do? change the memory-modules?

--
mezi@nkkorhaz.hu
2 REPLIES 2
malvin drakley
Esteemed Contributor
Solution

Re: D390 memory error; change mem?

Hi if you do a search on "sbe" on the forum you will see lots of posts about your problem annd the general concensus is, that if you only have a few errors then leave it, because the machine does its own correction of single bit errors. If you start getting a lot more then it will need investigating but only 4 errors is not too bad
cheers
malvin
Not me Chief, I'm Radar
Bill Hassell
Honored Contributor

Re: D390 memory error; change mem?

The D-class will automatically deallocate addresses that have too many correctable errors. Since your PDT has only 4 entries, there is no cause for concern yet. If you are running the online diagnostics, root will receive email when another error occurs. Multiple correctable errors in a single day is a sign that you need to order some new memory modules. If the error corrections start occurring at a rapid rate, your system will slow down dramatically. The reason is that the kernel must stop all processors except one while the memory problem is handled (a PDT entry made). In a 4 processor system, the system overheaad can jump to 90% or more and everything including login will crawl.


Bill Hassell, sysadmin