Informix on Solaris Zones: The server ran out of memory during anattempt to allocate a new shared memory segment

Omar Muñoz

2013-11-28 14:53:43 UTC

Thank for your answer.

Project configuration seems to be ok. The output from an "informix" user session on the non-global zone prior to start the database is as follows:

JUPITER_V890 jupiter_ids:>id -p
uid=1001(informix) gid=101(informix) projid=200(IDS-ATLAS)
JUPITER_V890 jupiter_ids:>projects -l IDS-ATLAS
IDS-ATLAS
projid : 200
comment: ""
users : informix
groups : informix
attribs: process.max-sem-nsems=(priv,2048,deny)
project.max-sem-ids=(priv,1024,deny)
project.max-shm-ids=(priv,256,deny)
project.max-shm-memory=(priv,12884901888,deny)
project.max-msg-ids=(priv,128,deny)

With this, max-shm-memory was set to 12 Gb, since the zone has got 12 Gb fo physical memory assigned from 32 Gb of the entire server as done at capped-memory at zone's configuration

I've noticed that decresing SHMVIRTSIZE and BUFFERPOOL I was able to start up informix - done as a test, since I must preserve the original values - but I can't go beyond 4.74 Gb of allocated RAM on the instance as shown in "onstat - ". While doing so, I've noticed that prstat -Z command shown the zone as if all informix memory was sent to swap. For instance, this was the output with 1 Gb aprox.

PID USERNAME SIZE RSS STATE PRI NICE TIME CPU PROCESS/NLWP
21315 informix 3928K 3344K cpu2 59 0 0:00:00 1.1% prstat/1
20405 noaccess 163M 124M sleep 59 0 0:00:33 0.6% java/18
20783 informix 1104M 280M sleep 59 0 0:00:04 0.0% oninit/2
20466 informix 7680K 4944K sleep 59 0 0:00:00 0.0% sshd/1
19891 root 7000K 4176K sleep 59 0 0:00:00 0.0% nscd/29
20477 informix 1888K 1424K sleep 59 0 0:00:00 0.0% ksh/1
19826 root 11M 9880K sleep 59 0 0:00:11 0.0% svc.configd/18
20784 informix 1103M 14M sleep 59 0 0:00:00 0.0% oninit/1
20128 root 4256K 2048K sleep 59 0 0:00:00 0.0% syslogd/14
20119 root 6104K 2256K sleep 59 0 0:00:00 0.0% automountd/3
20785 informix 1103M 13M sleep 59 0 0:00:00 0.0% oninit/1
19970 root 6888K 4200K sleep 59 0 0:00:01 0.0% inetd/4
19974 root 2464K 1432K sleep 59 0 0:00:00 0.0% sac/1
20791 informix 1103M 13M sleep 59 0 0:00:00 0.0% oninit/1
20123 root 4632K 2040K sleep 59 0 0:00:00 0.0% sshd/1
20474 root 9368K 2792K sleep 59 0 0:00:00 0.0% sendmail/1
20788 informix 1103M 13M sleep 59 0 0:00:00 0.0% oninit/1
20160 root 2928K 1792K sleep 59 0 0:00:00 0.0% snmpdx/1

ZONEID NPROC SWAP RSS MEMORY TIME CPU ZONE
12 46 1324M 227M 1.9% 0:00:51 1.9% jupiter_ids

Is there any limit or configuration to fit related to swap memory? I hope this new information were useful. Thanks in advance

--------------------------------------------
On Wed, 11/27/13, Habichtsberg, Reinhard <***@arz-emmendingen.de> wrote:

Subject: RE: Informix on Solaris Zones: The server ran out of memory during anattempt to allocate a new shared memory segment
To: "Omar Muñoz" <***@yahoo.com>
Date: Wednesday, November 27, 2013, 5:30 AM

IIRC Solaris restricts the use of
memory to about 4 GB for one project per default.

To avoid this restriction you have to grant user informix
more memory.

The content of /etc/project here (in this case it's the
global zone):
system:0::::
user.root:1::::
noproject:2::::
default:3::::
group.staff:10::::
group.informix:100:Informix DB
Umgebung:informix::project.max-sem-ids=(privileged,4096,deny);project.max-shm-ids=(privileged,256,deny);project.max-shm-memory=(priv,34359738368,deny)

The output from projects -l
...
group.informix
projid : 100
comment: "Informix DB Umgebung"
users : informix
groups : (none)
attribs:
project.max-sem-ids=(privileged,4096,deny)

project.max-shm-ids=(privileged,256,deny)

project.max-shm-memory=(priv,34359738368,deny)

AFAIK you can use the commands projadd, projmod and projdel
to configure project in a local zone, too. It's important
that the user who starts the oninits (informix) belongs to
the project-group with advanced ressource usage
(here:informix). Alternativly you can edit the
/etc/project in the local zone and restart the zone.

Please let me know if it helps.

Thanks, Reinhard.

-----Original Message-----

[mailto:informix-list-***@iiug.org]
On

Behalf Of Omar Muñoz
Sent: Tuesday, November 26, 2013 10:35 PM
Subject: Informix on Solaris Zones: The server ran out

of memory during

anattempt to allocate a new shared memory segment
Hi.
I'm installing informix 11.70.FC6 on a non-global

Solaris Zone (sparse root zone)

on a V890 machine Solaris 10. I can put online an

instance with 1Gb of RAM, but

when I tried to start a 8 Gb instance - alone or with

the other one online - it failed

and wrote in the online.log
14:23:38 IBM Informix Dynamic Server Started.
14:23:38 shmget: [ENOMEM][12]: key 52564801: The

server ran out of memory

during an attempt to allocate a new shared memory

segment.

14:23:38 mt_shm_init: can't create resident

segment

Memory parameters are
RESIDENT 0
SHMBASE 0x10a000000
SHMVIRTSIZE 2048000
SHMADD 40000
EXTSHMADD 8192
SHMTOTAL 0
SHMVIRT_ALLOCSEG 0.000000
SHMNOACCESS
Pools
BUFFERPOOL

default,buffers=10000,lrus=8,lru_min_dirty=50.00,lru_max_dirty=60.50

BUFFERPOOL

size=2K,buffers=2000000,lrus=16,lru_min_dirty=50.00,lru_max_dirty=60.0

0
BUFFERPOOL

size=8K,buffers=10000,lrus=8,lru_min_dirty=50.00,lru_max_dirty=60.50

Hardware seems to be enough: 32 Gb RAM. We're able to

start same instance on

global zone.
Zone configuration seems to be right too, but since I'm

new in virtualization, I'm

not so sure it's the proper configuration to run
zonecfg -z jupiter_ids info
zonename: jupiter_ids
zonepath: /jupiter_ids
brand: native
autoboot: true
ip-type: shared
[max-shm-memory: 12G]
    dir: /lib
    dir: /platform
    dir: /sbin
    dir: /usr
    dir: /export/home0
    special: /dev/dsk/c1t5d0s1
    raw: /dev/rdsk/c1t5d0s1
    type: ufs
    options: []
    dir: /respaldo2
    special: /dev/dsk/c1t5d0s3
    raw: /dev/rdsk/c1t5d0s3
    type: ufs
    options: []
    dir: /devices
    special: /devices
    raw not specified
    type: lofs
    options: []
    address: 192.168.33.2
    physical: ce0
    defrouter: 192.168.33.252
device
    match: /dev/rmt/*
device
    match: /dev/md/dsk/*
device
    match: /dev/md/rdsk/*
device
    match: /devices/pseudo/*
device
    match: /dev/dsk/*
device
    match: /dev/rdsk/*
    ncpus: 1
    physical: 12G
    [swap: 5G]
    [locked: 3G]
    name: zone.max-swap

(priv=privileged,limit=5368709120,action=deny)

name: zone.max-locked-memory

(priv=privileged,limit=3221225472,action=deny)

name: zone.max-shm-memory

(priv=privileged,limit=12884901888,action=deny)

Once into the zone, I checked prctl of the session from

which I tried to start up

informix
JUPITER_V890 jupiter_ids:>ps -o taskid -p $$ TASKID
3139
JUPITER_V890 jupiter_ids:>prctl -i task 3139
task: 3139
NAME PRIVILEGE

   VALUE
FLAG   ACTION

   RECIPIENT

task.max-cpu-time
system

18.4Es
inf none

-

task.max-lwps
system

2.15G
max deny

-

project.max-contracts
privileged

10.0K
- deny

-

project.max-device-locked-memory
privileged

1.96GB
- deny

-

project.max-locked-memory
system

16.0EB
max deny

-

project.max-port-ids
privileged

8.19K
- deny

-

project.max-shm-memory
privileged

12.0GB
- deny

-

project.max-shm-ids
privileged

256
- deny

-

project.max-msg-ids
privileged

128
- deny

-

project.max-sem-ids
privileged

1.02K
- deny

-

project.max-crypto-memory
privileged

7.82GB
- deny

-

project.max-tasks
system

2.15G
max deny

-

project.max-lwps
system

2.15G
max deny

-

project.cpu-cap
system

4.29G
inf deny

-

project.cpu-shares
privileged

1
- none

-

zone.max-swap
privileged

5.00GB
- deny

-

zone.max-locked-memory
privileged

3.00GB
- deny

-

zone.max-shm-memory
privileged

12.0GB
- deny

-

zone.max-shm-ids
system

16.8M
max deny

-

zone.max-sem-ids
system

16.8M
max deny

-

zone.max-msg-ids
system

16.8M
max deny

-

zone.max-lwps
system

2.15G
max deny

-

zone.cpu-cap
system

4.29G
inf deny

-

zone.cpu-shares
privileged

1
- none

-

Any suggestions on how to overcome this

would be appreciated.

_______________________________________________
Informix-list mailing list
http://www.iiug.org/mailman/listinfo/informix-list