topic Re: mmap failed in Operating System - HP-UX

mmap failed

brian_31 — Fri, 09 Mar 2007 10:20:07 GMT

# ./java --version
/usr/lib/dld.sl: Call to mmap() failed - TEXT /opt/java1.4/jre/lib/PA_RISC2.0/server/libjvm.sl
/usr/lib/dld.sl: Not enough space
Abort(coredump)

Can someone help with this error?

Thanks

brian

Re: mmap failed

brian_31 — Fri, 09 Mar 2007 10:30:02 GMT

# swapinfo -atm
Mb Mb Mb PCT START/ Mb
TYPE AVAIL USED FREE USED LIMIT RESERVE PRI NAME
dev 16384 10758 5626 66% 0 - 1 /dev/vg00/lvol2
reserve - 4212 -4212
memory 6184 3811 2373 62%
total 22568 18781 3787 83% - 0 -

swap looks ok too

Thanks

brian

Re: mmap failed

Sundar_7 — Fri, 09 Mar 2007 11:10:50 GMT

hmm, I would be really surprised if your max data and text segment sizes are too small. But doesn't hurt to check.

# kmtune | egrep "dsiz|tsiz"

Post what you get.

Re: mmap failed

brian_31 — Fri, 09 Mar 2007 11:56:11 GMT

# kmtune | egrep "dsiz|tsiz"
maxdsiz 1073741824 - 1073741824
maxdsiz_64bit 0x40000000 - 0X40000000
maxtsiz 0x40000000 Y 0X40000000
maxtsiz_64bit 0x40000000 Y 0X40000000
#

Brian

Re: mmap failed

Sundar_7 — Fri, 09 Mar 2007 16:05:24 GMT

Your dsize and tsize seems to be good enough.

For the lack of other responses, I can throw some wild guesses :-).

Is this Java by HP or some other vendor ? In other words, did you install the product using swinstall ? If yes, you could try a swverify to make sure you dont have a corrupt .sl.

Re: mmap failed

Dennis Handly — Fri, 09 Mar 2007 18:12:33 GMT

Since your message says "Call to mmap() failed - TEXT", you are short on shared memory, not swap space.

What does the following show:
$ ipcs -ma

Any large segments?

Re: mmap failed

sala.trancingular.com — Mon, 19 Nov 2007 13:56:08 GMT

I have the same problem here. Would you able to help.
# swapinfo -t
Kb Kb Kb PCT START/ Kb
TYPE AVAIL USED FREE USED LIMIT RESERVE PRI NAME
dev 4194304 0 4194304 0% 0 - 1 /dev/vg00/lvol3
dev 10240000 0 10240000 0% 0 - 1 /dev/vg00/lvol7
dev 5242880 0 5242880 0% 0 - 1 /dev/vg00/lvol8
dev 10240000 0 10240000 0% 0 - 1 /dev/vg00/lvol14
dev 4194304 0 4194304 0% 0 - 1 /dev/vg00/lvol15
reserve - 34111488 -34111488
memory 100119552 49524696 50594856 49%
total 134231040 83636184 50594856 62% - 0 -

# kctune |egrep "dsiz|tsiz"
maxdsiz 570425344 0X22000000 Immed
maxdsiz_64bit 1073741824 0X40000000 Immed
maxtsiz 0x8000000 0x8000000 Immed
maxtsiz_64bit 0x40000000 0x40000000 Immed

# ipcs -ma
IPC status from /dev/kmem as of Mon Nov 19 10:55:48 2007
T ID KEY MODE OWNER GROUP CREATOR CGROUP NATTCH SEGSZ CPID LPID ATIME DTIME CTIME
Shared Memory:
m 0 0x41302a7f --rw-rw-rw- root root root root 0 348 1761 28370 20:06:22 20:06:22 7:14:23
m 1 0x4e100002 --rw-rw-rw- root root root root 2 61760 1761 28370 20:06:22 20:06:22 7:14:24
m 2 0x412c2340 --rw-rw-rw- root root root root 2 8192 1761 28370 20:06:22 20:06:22 7:14:24
m 3 0x0c6629c9 --rw-r----- root edasql root edasql 2 43963752 3248 9036 10:54:18 10:54:27 7:15:43
m 4 0x06347849 --rw-rw-rw- root edasql root edasql 2 65544 3248 6411 0:26:03 7:21:56 7:15:43
m 5 0x49340041 --rw-r--r-- root edasql root edasql 0 22908 3247 4125 10:55:01 10:55:01 7:15:44
m 294918 0x00005643 --rw-rw-rw- root root root root 1 1024 3489 23290 7:21:56 7:21:56 7:16:10
m 7 0x00005654 --rw-rw-rw- root root root root 1 1024 3489 23290 7:21:56 7:21:56 7:16:10
m 5177352 0x741cc1a0 --rw-rw-rw- root root root root 0 28244 6812 23552 10:55:09 10:55:09 7:18:19
m 3473417 0x632dca48 --rw-r----- oracle dba oracle dba 2223 13941895168 18211 7797 10:55:39 10:55:39 22:46:30
m 32778 0x7fffff00 --rw------- root root root root 0 49277 4920 4920 7:16:49 no-entry 7:16:49
m 491531 0x2019bf20 --rw-r----- oracle dba oracle dba 30 3774885888 18505 7702 10:55:39 10:55:39 22:46:56

Re: mmap failed

Bill Hassell — Mon, 19 Nov 2007 21:18:48 GMT

Your Java code is probably 32bit and is competing with all other 32bit processes for the single shared memory map. While there may be enough in the shared memory map for all your processes, it has become fragmented and none of the pieces is large enough. This usually comes from randomly shutting down various processes or worse, using kill -9. A reboot will likely fix the problem, or possibly, a graceful shutdown of the database and applications, check ipcs -bmop to see if there are any orphaned segments. Then start the processes again.