Sys-Admin Information's » Solaris

Boot archive error or corrupt

Hendrawan — Fri, 01 May 2009 06:50:45 +0000

A few weeks ago, I got a little problem with my solaris box, namely solaris won't boot because the boot archive was corrupt, *sigh*. Anyway the boot archive in solaris 10 was contain kernel module and configuration file was needed for solaris to startup the system.

Error:

module /platform/i86pc/boot_archive error 3 bad or corrupt data while decompressing file

Workaround:

Boot up your solaris in "solaris failsafe mode", next solaris image will mount with writeable mode on "/a" mount point

sh> rm -f /a/platform/i86pc/boot_archive
sh> bootadm update-archive -R /a
sh> reboot

IP multipathing on Solaris 10 with Hanic (NIC Redundancy with one IP)

Hendrawan — Sat, 25 Apr 2009 09:20:55 +0000

Hanic ?
Do you know what's that.
Yes, Hanic is NIC Redundancy with one IP, hanic was made by "Partha Subramanian", and based on perl, hanic more simply and powerfull than IPMP, the default solaris IP multipathing. Why hanic was simply, in hanic we unnecessary to defined all requirement for respective NIC , just install, run and ready to failover.

Bellow, this is the step by step for hanic used.

1. First of all, we must got the hanic source as usually

bash> wget http://www.sun.com/bigadmin/scripts/submittedScripts/hanic.tar.gz

2. Show your NIC for me

bash> ifconfig e1000g0 plumb
bash> ifconifg e1000g1 plumb

bash> vi /etc/hosts
192.168.85.3    training.bascomp.org    training

3. Install hanic on your system

bash> gunzip hanic.tar.gz|tar xvf -
bash> cd hanic
bash> chmod +x *
bash> ./hanic.install

As usually software installation you must follow the guide, something like this.
Define your primary and secondary NIC, type your IP Address, your netmask and then reboot your system, for completed the installation.

bash> init 6

4. Failover IP Address

bash> hanic -monitor;hanic -cleanup
bash> hanic -status
 
STATUS - HA NIC Monitoring
-------------------------------------------------
State       Active(Passive)    PID
-------------------------------------------------
Running     e1000g1(e1000g0)   1081
-------------------------------------------------

Oke. seem to be hanic work as predictable, furthermore we test the fail over with "switch" command, whereabout "switch" do the switch ip address from primary NIC to secondary NIC.

bash> ifconfig -a
e1000g1: flags=1000843&lt;UP,BROADCAST,RUNNING,MULTICAST,IPv4&gt; mtu 1500 index 6
inet 192.168.85.3 netmask ffffff00 broadcast 192.168.85.255
ether 8:0:27:86:2e:c4

bash> hanic -status
 
STATUS - HA NIC Monitoring
-------------------------------------------------
State       Active(Passive)    PID
-------------------------------------------------
Running     e1000g1(e1000g0)   1081
-------------------------------------------------

bash> hanic -switch
 
bash> hanic -status
 
STATUS - HA NIC Monitoring
-------------------------------------------------
State       Active(Passive)    PID
-------------------------------------------------
Running     e1000g0(e1000g1)   1081
-------------------------------------------------
 
bash> ifconfig -a
e1000g0: flags=1000843&lt;UP,BROADCAST,RUNNING,MULTICAST,IPv4&gt; mtu 1500 index 7
inet 192.168.85.3 netmask ffffff00 broadcast 192.168.85.255
ether 8:0:27:82:21:80

5. Failover offline interface
Oke works perfect ha, now lets start second trial, for this situation try to offline the active network interface, and see what's going, for monitoring this trial, you can see wheter the connection drop or not with ping command.

bash> ifconfig -a
e1000g0: flags=1000843&lt;UP,BROADCAST,RUNNING,MULTICAST,IPv4&gt; mtu 1500 index 3
inet 192.168.85.3 netmask ffffff00 broadcast 192.168.85.255
ether 8:0:27:82:21:80
 
bash> hanic -offline
 
bash> ifconfig -a
e1000g1: flags=1000843&lt;UP,BROADCAST,RUNNING,MULTICAST,IPv4&gt; mtu 1500 index 4
inet 192.168.85.3 netmask ffffff00 broadcast 192.168.85.255
ether 8:0:27:86:2e:c4
 
bash> hanic -status
 
STATUS - HA NIC Monitoring
-------------------------------------------------
State       Active(Passive)    PID
-------------------------------------------------
Running     e1000g1(e1000g0)   1081
-------------------------------------------------

Source from: http://dot.choosen.net/2009/04/06/ip-multipathing-on-solaris-10-with-hanic-nic-redundancy-with-one-ip/

Mirroring/Unmirroring on Solaris 10

Hendrawan — Sat, 25 Apr 2009 07:49:43 +0000

Post kali ini dengan bahasa indonesia tentunya, akan membahas mengenai membuat mirror dan melepaskan mirror di solaris 10 x86.
sekedar informasi, RAID-1 yang dikenal dengan mirror ini sangat populer untuk tingkat data redudansi juga faster writeable performance,
RAID-1 biasanya digunakan untuk system operasi ditempatkan, dan juga untuk aplikasi dengan tingkat penulisan yang tinggi, contoh Redolog.
Memang sangat di yakinin bahwa untuk pembahasan mengenai RAID-1 ini sangat-sangat banyak bertebaran di internet, tapi apalah daya untuk kepentingan dokumentasi kita bahas disini.
Postingan RAID-1 ini menggunakan solaris volume manager (SVM), dan 2 disk dengan size yang sama.

Magic code here.
Dibawah ini adalah kondisi awal sebelum 2 disk baru diinstallasi di system.

sh> cfgadm -al
Ap_Id                          Type         Receptacle   Occupant     Condition
sata0/0::dsk/c1t0d0            disk         connected    configured   ok

Kita lanjut untuk rekonfigur boot proses dan updating /dev dan /devices.

sh> devfsadm -C
sh> cfgadm -al
Ap_Id                          Type         Receptacle   Occupant     Condition
sata0/0::dsk/c1t0d0            disk         connected    configured   ok
sata0/1::dsk/c1t1d0            disk         connected    configured   ok
sata0/2::dsk/c1t2d0            disk         connected    configured   ok

Oke setelah semua persiapan selesai, kita lanjut dengan membuat RAID-1nya.

sh> format c1t1d0

lalu buat 2 partisi, satu untuk data dan satulagi untuk metadatanya disimpan, biasanya di slice 7.
Setelahnya kita salin format partisi dari disk primary ke disk secondary, dan dilanjut dengan membuat state databasenya.

sh> prtvtoc /dev/rdsk/c1t1d0s2 |fmthard -s - /dev/rdsk/c1t2d0s2
sh> metadb -a -f -c 3 c1t1d0s7 c1t2d0s7

sh> metadb
        flags           first blk       block count
     a m  pc luo        16              8192            /dev/dsk/c1t1d0s7
     a    pc luo        8208            8192            /dev/dsk/c1t1d0s7
     a    pc luo        16400           8192            /dev/dsk/c1t1d0s7
     a    pc luo        16              8192            /dev/dsk/c1t2d0s7
     a    pc luo        8208            8192            /dev/dsk/c1t2d0s7
     a    pc luo        16400           8192            /dev/dsk/c1t2d0s7

Setelah konfirm, lanjut dengan membuat raid 1 nya.

sh> metainit -f d70 1 1 c1t1d0s3
sh> metainit -f d71 1 1 c1t2d0s3
sh> metainit d69 -m d70
sh> metattach d69 d71

sh> metastat d69
d69: Mirror
    Submirror 0: d70
      State: Okay         
    Submirror 1: d71
      State: Resyncing    
    Resync in progress: 35 % done
    Pass: 1
    Read option: roundrobin (default)
    Write option: parallel (default)
    Size: 4194304 blocks (2.0 GB)
 
d70: Submirror of d69
    State: Okay         
    Size: 4194304 blocks (2.0 GB)
    Stripe 0:
        Device     Start Block  Dbase        State Reloc Hot Spare
        c1t1d0s3          0     No            Okay   Yes 
 
 
d71: Submirror of d69
    State: Resyncing    
    Size: 4194304 blocks (2.0 GB)
    Stripe 0:
        Device     Start Block  Dbase        State Reloc Hot Spare
        c1t2d0s3          0     No            Okay   Yes

Terlihat sedang dalam proses resync (Resync in progress: 35 % done),
kita lanjut dengan membuat mount pointnya.

sh> newfs -f 4096 /dev/md/rdsk/d69
sh> mkdir /apps
sh> vi /etc/vfstab
/dev/md/dsk/d69 /dev/md/rdsk/d69 /apps ufs 2 yes -

shell> mountall
shell> df -F ufs -h|grep apps
/dev/md/dsk/d69        1.9G   2.0M   1.9G     1%    /apps

Oke lanjut dengan Unmirroringnya, biar komplit.
Lakukan "metadetach" untuk membuat one-way mirror.

sh> umount /apps
sh> metadetach d69 d71
d69: submirror d71 is detached

Verfiy...

sh> metastat d69
d69: Mirror
    Submirror 0: d70
      State: Okay         
    Pass: 1
    Read option: roundrobin (default)
    Write option: parallel (default)
    Size: 4194304 blocks (2.0 GB)
 
d70: Submirror of d69
    State: Okay         
    Size: 4194304 blocks (2.0 GB)
    Stripe 0:
        Device     Start Block  Dbase        State Reloc Hot Spare

Tahap akhir adalah, "metaclear" untuk meng-clear mirror dan submirrornya.

sh> metaclear -r d69
d69: Mirror is cleared
d70: Concat/Stripe is cleared
sh> metaclear d71
d71: Concat/Stripe is cleared

Jangan lupa untuk menghilangkan "/apps" dari /etc/vfstabnya.

PowerPath show "unlic" status

Hendrawan — Mon, 23 Mar 2009 13:02:32 +0000

Unlic indicates that unlicensed PowerPath is running for a CLARiiON array (no license key has been installed), Unlicensed defaults to "basic Fail/Over policy, we made change yet.

bash> powermt display dev=emcpower0a|grep policy
state=alive; policy=BasicFailover; priority=0; queued-IOs=0

bash> powermt display dev=emcpower0a
Pseudo name=emcpower0a
CLARiiON ID=CK200071500374 [SERVER]
Logical device ID=60060160F0CE1A00185D33F95722DD11 [LUN 25 SERVER]
state=alive; policy=BasicFailover; priority=0; queued-IOs=0
Owner: default=SP B, current=SP B       Array failover mode: 1
==============================================================================
---------------- Host ---------------   - Stor -   -- I/O Path -  -- Stats ---
###  HW Path                I/O Paths    Interf.   Mode    State  Q-IOs Errors
==============================================================================
3072 pci@8,700000/SUNW,qlc@3/fp@0,0 c5t5006016239A02677d0s0 SP A2     active  alive      0      0
3072 pci@8,700000/SUNW,qlc@3/fp@0,0 c5t5006016A39A02677d0s0 SP B2     active  alive      0      0
3074 pci@8,700000/SUNW,qlc@3,1/fp@0,0 c6t5006016339A02677d0s0 SP A3     unlic   alive      0      0
3074 pci@8,700000/SUNW,qlc@3,1/fp@0,0 c6t5006016B39A02677d0s0 SP B3     unlic   alive      0      0

Workaround
1. powermt check_registration
2. powermt set policy=co dev=all (Set the Policy to co = Clariion Optimization)
3. powermt display dev=all
4. powermt save

bash> powermt display dev=emcpower0a|grep policy
state=alive; policy=CLAROpt; priority=0; queued-IOs=0

bash> powermt display dev=emcpower0a
Pseudo name=emcpower0a
CLARiiON ID=CK200071500374 [SERVER]
Logical device ID=60060160F0CE1A00185D33F95722DD11 [LUN 25 SERVER]
state=alive; policy=CLAROpt; priority=0; queued-IOs=0
Owner: default=SP B, current=SP B       Array failover mode: 1
==============================================================================
---------------- Host ---------------   - Stor -   -- I/O Path -  -- Stats ---
###  HW Path                I/O Paths    Interf.   Mode    State  Q-IOs Errors
==============================================================================
3072 pci@8,700000/SUNW,qlc@3/fp@0,0 c5t5006016239A02677d0s0 SP A2     active  alive      0      0
3072 pci@8,700000/SUNW,qlc@3/fp@0,0 c5t5006016A39A02677d0s0 SP B2     active  alive      0      0
3074 pci@8,700000/SUNW,qlc@3,1/fp@0,0 c6t5006016339A02677d0s0 SP A3     active  alive      0      0
3074 pci@8,700000/SUNW,qlc@3,1/fp@0,0 c6t5006016B39A02677d0s0 SP B3     active  alive      0      0

Set hostname without reboot

Gun nee — Mon, 28 Apr 2008 07:00:46 +0000

bash> uname -a
SunOS Trojan 5.9 Generic_118558-33 sun4u sparc SUNW,Netra-T12

replace old "hostname" with new "hostname" in file /etc/hosts, /etc/nodename, /etc/hostname."ethernet_port_name", /etc/net/ticlts/hosts, /etc/net/ticotsord/hosts, /etc/net/ticots/hosts

bash> vi /etc/hosts
bash> vi /etc/nodename
bash> vi /etc/hostname.bge0
bash> vi /etc/net/ticlts/hosts
bash> vi /etc/net/ticotsord/hosts
bash> vi /etc/net/ticots/hosts

and then execute command :

hostname "new hostname"

bash> hostname worm
bash> . /.profile

ssh[23593]: [ID 530472 user.error] Kerberos mechanism library initialization error:

Hendrawan — Mon, 24 Mar 2008 02:29:31 +0000

When I log via ssh to another machine

bash> ssh gandalf@rohan.lotr.com
unable to initialize mechanism library [/usr/lib/gss/gl/mech_krb5.so]

the following message appears was appears

rohan.lotr.com sshd[23418]: [ID 530472 auth.error] Kerberos mechanism library initialization error:
Can't open/find Kerberos /etc/krb5/krb5.conf

Workaround:

Edit /etc/ssh/ssh_config and add the following

Host *
StrictHostKeyChecking no
GSSAPIKeyExchange no
GSSAPIAuthentication no

And
Edit /etc/ssh/sshd_config and add the following

GSSAPIAuthentication no
GSSAPIKeyExchange no
GSSAPIStoreDelegatedCredentials no

Comparison Kernel parameter Solaris9 and Solaris10

Hell-me — Tue, 04 Mar 2008 11:27:12 +0000

SERVER - Solaris9 - Oracle9i .....[compare]..... SERVER - solaris10 - Oracle10g

[ Solaris9] forceload:sys/shmsys
[Solaris10] recommended not set, server impact long time for boot up

[ Solaris9] forceload:sys/semsys
[Solaris10]recommended not set, server impact long time for boot up

[ Solaris9] set shmsys:shminfo_shmmax=51539607552
[Solaris10] projmod -s -K "project.max-shm-memory=(priv,48g,deny)" user.oracle

[ Solaris9] set shmsys:shminfo_shmmin=1
[Solaris10] Obsolete in the Solaris 9 release, Solaris 10 also

[ Solaris9] set shmsys:shminfo_shmmni=2048
[Solaris10] projmod -s -K "project.max-shm-ids=(priv,2048,deny)" user.oracle

[ Solaris9] set shmsys:shminfo_shmseg=200
[Solaris10] Obsolete in the Solaris 9 release, Solaris 10 also

[ Solaris9] set semsys:seminfo_semmni=1000
[Solaris10] projmod -s -K "project.max-sem-ids=(priv,1000,deny)" user.oracle

[ Solaris9] set semsys:seminfo_semmns=32767
[Solaris10] Removed in the Solaris 10 release

[ Solaris9] semsys:seminfo_semmnu=500
[Solaris10] Removed in the Solaris 10 release.

[ Solaris9] set semsys:seminfo_semmsl=2048.
[Solaris10] projmod -s -K "project.max-sem-nsems=(priv,4096,deny)" user.oracle

[ Solaris9] semsys:seminfo_semvmx=32767.
[Solaris10] Removed in the Solaris 10 release

[ Solaris9] set semsys:seminfo_semopm=2048
[Solaris10] projmod -s -K "process.max-sem-ops=(priv,2048,deny)" user.oracle

[ Solaris9] set maxusers=2048
[Solaris10] recommended not set

[ Solaris9] set max_nprocs=32778
[Solaris10] set max_nprocs=32778

[ Solaris9] set maxuprc=32773
[Solaris10] set maxuprc=32773

[ Solaris9] set rlim_fd_max=65536
[Solaris10] set rlim_fd_max=65536

[ Solaris9] set rlim_fd_cur=2048
[Solaris10] set rlim_fd_cur=2048

[ Solaris9] set noexec_user_stack=1
[Solaris10] set noexec_user_stack=1

*For Veritas NetBackup Performance
[ Solaris9] set msgsys:msginfo_msgtql=512
[Solaris10] Obsolete in the Solaris 10 release

[ Solaris9] set msgsys:msginfo_msgmnb=65536
[Solaris10] Obsolete in the Solaris 10 release

---------------------------------------------------------------------------------------------------------------------------------------
Conclusion ,the following are the important parameters:

*project.max-shm-ids: Maximum shared memory IDs for a project. Replaces shmmni

*project.max-sem-ids: Maximum semaphore IDs for a project.Replaces semmni

*project.max-msg-ids: Maximum message queue IDs for a project. Replaces msgmni

*project.max-shm-memory: Total amount of shared memory allowed for a project. Replaces shmmax

*process.max-sem-nsems: Maximum number of semaphores allowed per semaphore set.Replaces semmsl

*process.max-sem-ops: Maximum number of semaphore operations allowed per semop.Replaces semopm

*process.max-msg-qbytes: Maximum number of bytes of messages on a message queue. Replaces msgmnb

*process.max-msg-messages: Maximum number of messages on a message queue.Replaces msgtql

Kernel setup for Solaris 10 using project files - Special for Oracle10g

Hell-me — Tue, 04 Mar 2008 10:20:49 +0000

1. Recommended method for modifying the /etc/project file is to use the "projadd" and "projmod" :

bash> projadd -c "Oracle" 'user.oracle'

bash> projmod -s -K "project.max-shm-memory=(privileged,10GB,deny)" 'user.oracle'

c. verify with

bash> cat /etc/project

bash> prctl -i project user.oracle

2. Resource Control assignments made in this way (in the /etc/project file) are permanent, and will survive a system re-boot.

3. There is also an "on-the-fly" way to temporarily set Resource Control assignments using the prctl(1) command.

a. Should we need to temporarily increase the setting to 48 GB:
For example, assuming the preceding /etc/project

bash> prctl -n project.max-shm-memory -i project user.oracle
project: 100: user.oracle
NAME PRIVILEGE VALUE FLAG ACTION RECIPIENT
project.max-shm-memory
privileged 16.0GB - deny -
system 16.0EB max deny -

bash> prctl -n project.max-shm-memory -r -v 48GB -i project user.oracle

bash> prctl -n project.max-shm-memory -i project user.oracle
project: 100: user.oracle
NAME PRIVILEGE VALUE FLAG ACTION RECIPIENT
project.max-shm-memory
privileged <strong>48strong> .0GB - deny -
system 16.0EB max deny

b. fix parameter ; entry new value on max-shm

bash> vi /etc/project

values such as 48GB instead of the 51539607552

project.max-shm-memory=(privileged, 51539607552, deny);