Operating System - Tru64 Unix
1748201 Members
2983 Online
108759 Solutions
New Discussion юеВ

Re: signal 6- resource lost

 
Tong Wang
New Member

signal 6- resource lost

Hi all,

Hi all,
I run my application on tru64(OSF) machine. In my application, when loading data from database recently after we add more data, it gives the signal 6,resource lost. The following is basic info about my system and the process. Can someone tell me what's
the criteria to issue the signal in the system?

If I remove some of the data from database, the application won't get any signal.
I noticed that the application takes about 25.5% percent of the phyiical memory(pmem)when it gets the signal. The physical memory of my system is 1GB. is there any limit for the one process to take up the physical memory.

Thank you for any comments and suggestions.

please copy to me: tong_wang@efunds.com when you reply.

Tong

Physical Memory Use: (by vmstat -P)
start_pfn end_pfn type size_pages / size_bytes
256 288 scavenge 32 / 256.00k
288 1023 text 735 / 5.74M
1023 1158 data 135 / 1.05M
1158 1391 bss 233 / 1.82M
1391 1398 cfgmgmt 7 / 56.00k
1398 1400 locks 2 / 16.00k
1400 1414 pmap 14 / 112.00k
1414 2033 unixtable 619 / 4.84M
2033 2057 logs 24 / 192.00k
2057 4105 vmtables 2048 / 16.00M
4105 130985 managed 126880 / 991.25M
============================
Total Physical Memory Use: 130729 / 1021.32M
---------------------------------------------------------------
PID RSS VSZ %MEM SCNT PENDING UPROCP S TTY TIME
COMMAND
498647 584K 2.39M 0.1 0 0 212cc0 S pts/0 0:00.52
ksh
498666 592K 2.39M 0.1 0 0 bb2240 I + pts/1 0:00.20
ksh
499593 261M 375M 25.5 0 0 398cc0 U N pts/1 0:39.09
clearing
---------------------------------------------------------------
load averages: 0.08, 0.07, 0.06
12:57:17
64 processes: 1 running, 25 sleeping, 37 idle, 1 zombie
CPU states: 0.1% user, 0.0% nice, 1.0% system, 98.7% idle
Memory: Real: 234M/991M act/tot Virtual: 1M/3072M use/tot Free: 411M
PID USERNAME PRI NICE SIZE RES STATE TIME CPU COMMAND
689 root 44 0 5864K 2080K sleep 2:06 0.20% envmond
499823 wtong 44 0 5872K 3235K run 0:00 0.10% top
653 httpd 42 0 141M 134M sleep 25:31 0.00% java
499758 wtong 42 4 299M 172M sleep 0:31 0.00% clearing
978 root 42 0 19M 11M sleep 0:18 0.00% smsd
6 REPLIES 6
Vladimir Fabecic
Honored Contributor

Re: signal 6- resource lost

Did you take a look at ulimit parameters?
Maybe you need bigger memory quota.
Do man sys_attrs_proc.
In vino veritas, in VMS cluster
Muthukumar_5
Honored Contributor

Re: signal 6- resource lost

Your application with new data is having problem to proceed further. That is, program detected an illegal state and has aborted allowing for debugging.

Is there any related core file formed there. If so debug that with application binary. Check the new added data also.

hth.
Easy to suggest when don't know about the problem!
Ralf Puchner
Honored Contributor

Re: signal 6- resource lost

depending on the database kernel parameter tunings are required (e.g. heap, shared memory, stacksizes etc.)

So please have a look into the config guide of your database.
Help() { FirstReadManual(urgently); Go_to_it;; }
Tong Wang
New Member

Re: signal 6- resource lost

Thank you for replying. We tested the application on our client machine (tru64)with 8GB memory. It runs well without any problem as before. We run the application on other platform, AIX, HP, Solar, which have at least 2GB, without any problem.

We did get the core file. It seems not much help to me.

When we run on our machine(1GB memeory), we shut down all unnecessary applications, no other terminals connect to this machine.

We did not run with the debuggor. The application stop printting out message on terminal, and it takes several ten seconds to exit. During the period, we have enough time to run top etc.

Please kindly give me the hints what doc I should check and where can I find it. I am not sys adamin.

regards,

Tong
Ralf Puchner
Honored Contributor

Re: signal 6- resource lost

be sure system was properly configured (kernel parameters) for the database!
Help() { FirstReadManual(urgently); Go_to_it;; }
Tong Wang
New Member

Re: signal 6- resource lost

The database is oracle 9.2.1 which is running on the other machine which is different than that of my application running on.

I've checked out that the stack size is setup as:

per_proc_stack_size = 33554432

Can someone tell me how to check out the stack size in a process? I got the core file.

Thanks & regards,

Tong