errors at sisp start

Submitted by jozsefbiro on Mon, 2006-10-16 03:58.

Hi,

I installed the latest OpenClovis release and tried first to use the evaluation kit as described below:

1) I first started the evaluation wizard as instructed by 4.1 of the Evaluation System User Guide (i.e. I have a single node configuration that conforms to Runtime Setup 2.1 with the restriction that the Management Station and the System Controller is the same machine).

2) Then I jumped to 5.2.1 and started ASP:

[root@mymachine ~]# sisp/etc/init.d/sisp start
Starting sisp_amf: [ OK ]

3) Then I tried to run csa101 by jumping to 5.4.4, but this step did not go as expected:

[root@mymachine bin]# ./debug.sh

To get started, type 'help intro'

cli[Test]-> setc 1

CPM is not running in the given slot

Any help on what went wrong and how I could fix it would be appreciated!

BR Jozsef

Some other info that may be useful:

1: The Linux version I am running:
[root@mymachine bin]# uname -a
Linux mymachine 2.6.9-42.EL #1 Wed Jul 12 23:16:43 EDT 2006 i686 i686 i386 GNU/Linux

2: ps output after starting sisp, i.e. after step 2):

[root@mymachine ~]# ps aux | grep sisp
root 5073 0.2 0.3 9604 3624 pts/9 S 12:28 0:00 /root/sisp/bin/snmpd -DH -Lo -f -c /root/sisp/etc/snmpd.conf
root 5096 0.0 0.0 4968 652 pts/9 S+ 12:29 0:00 grep sisp

(Based on another post, I think that sisp should have started some other processes as well...)

3: sisp log after starting sisp, i.e. after step 2):

[root@mymachine bin]# tail -f /var/log/sisp

PID[5074]:File[cpm/clCpmBoot.c]:Func[cpmBmStartup]:Line[520]: Unable to reach default boot level [2], rc=[0xf0014]

PID[5074]:File[cpm/clCpmBoot.c]:Func[bmInitialize]:Line[179]: Booting up process failure f0014, not aborting at this point

PID[5074]:File[clLogClient.c]:Func[clLogLibFinalize]:Line[960]: Log Library not initialized by process 5074

PID[5074]:File[cpm/clCpmGms.c]:Func[cpmGmsFinalize]:Line[813]: Unable to delete mutex, rc=[0x1010a]

(And this log also seems to indicate some startup problem...)

Submitted by harikrishna_gp on Mon, 2006-10-16 17:08.

Hi ,
From the log messages you have posted it looks like Compoment manager(CPM) is
not able to start the Group Membership (GMS) process.
GMS process could fail to come up because of any of the following reasons.

1. Your target.conf has some IP addresses which are not correct
2. You already have SISP running with the same node number.
3. Your network does stop multicasting packets for some reason.

Please check these. Also if you see any more suspicious messages
in /var/log/sisp please post them. That can give more clues.

Thanks

Submitted by jozsefbiro on Wed, 2006-10-25 03:55.

Hi,

1) I checked the target.conf file and added the correct IP info, Here is the complete target.conf file (comments exluded):

TRAP_IP=127.0.0.1
CMM_IP=127.0.0.1
INSTALL_PREREQUISITES=YES
INSTANTIATE_IMAGES=YES
CREATE_TARBALLS=YES
LOG_ADDR=127.0.0.1
SCNODE_PATTERN='SCNodeI*'
IP_SCNodeI0=172.24.169.99
SLOT_SCNodeI0=1

Then, I cleaned the eval project and also /root/sisp and executed the steps in 4.1, 5.2.1 and finally 5.4.4 but I still get the same messages. Alternatively, I also tried to build target images manually (all steps 4.2 to 4.6, instead of 4.1) but that did not help either.

In fact, I am wondering how the IP address has any relevance in a single node configuration.

2) I did find that /var/log/sisp was more verbose than in my original post, though. Maybe that can give you more information on what is happenning here. So, here is sisp log after starting sisp, i.e. after 5.2.1:

[root@mymachine biro]# tail -f /var/log/sisp

PID[5049]:File[atca.c]:Func[openFd]:Line[480]: Unable to open device

PID[5049]:File[cpm/clCpmComponent.c]:Func[_cpmSaAwareComponentInstantiate]:Line[3275]: Image name : [sisp_logd], PID : [5060]

GMS:[info ] New Group Created with Group id :0
GMS:[info ] Waiting for Prospective Leaders...........PostPoning Leader Election for 5 secs

GMS:[info ] LeaderElection Initial Run
GMS:[debug] Server Got Termination Request Started Shutting Down
PID[5060]:File[clLogServerMain.c]:Func[clLogSvrTerminate]:Line[228]: Unregistering with cpm...... [logServer_SCNodeI0]

PID[5060]:File[clLogServerMain.c]:Func[clLogSvrFinalize]:Line[473]: Log Server Exitting Gracefully....

PID[5049]:File[cpm/clCpmComponent.c]:Func[_cpmSaAwareComponentInstantiate]:Line[3275]: Image name : [sisp_gms], PID : [5065]

PID[5049]:File[cpm/clCpmBoot.c]:Func[cpmBmSetLevel]:Line[1385]: Unable to set required boot level [2], rc=[0xf0014]

PID[5049]:File[cpm/clCpmBoot.c]:Func[cpmBmStartup]:Line[520]: Unable to reach default boot level [2], rc=[0xf0014]

PID[5049]:File[cpm/clCpmBoot.c]:Func[bmInitialize]:Line[179]: Booting up process failure f0014, not aborting at this point

PID[5049]:File[clLogClient.c]:Func[clLogLibFinalize]:Line[960]: Log Library notinitialized by process 5049

PID[5049]:File[cpm/clCpmGms.c]:Func[cpmGmsFinalize]:Line[813]: Unable to deletemutex, rc=[0x1010a]

[root@mymachine biro]#

Any hints on what could be wrong here will be highly appreciated!

BR

Jozsef

Submitted by harikrishna_gp on Wed, 2006-10-25 18:10.

Hi Jozsef,
This might sound like a silly clarification. There are multiple target.conf files that show up if you try to search for it.
I hope the target.conf file being modified is picked up from project-area_dir/SISP/models/eval/ directory.
Also knowing the versions of kernel,gcc,glibc,pthread library on ur machine would help alot.
BTW right now ASP supports RHEL4. Is the version of linux you are using different?

Thanks

Submitted by jozsefbiro on Wed, 2006-10-25 21:15.

Hi,

>> This might sound like a silly clarification. There are multiple target.conf files that show up if you try to search for it.
>> I hope the target.conf file being modified is picked up from project-area_dir/SISP/models/eval/ directory.

Yes, that's the one.

And yes, I know that there are more target.conf files. Here is the full list of them on my computer:

1) home/biro/clovis/SISP/build/common/conf/target.conf
2) home/biro/clovis/SISP/models/eval/target.conf
3) opt/clovis/sdk-2.2/src/SISP/build/common/conf/target.conf
4) opt/clovis/sdk-2.2/src/SISP/models/eval/target.conf

3) and 4) are in the OpenClovis installation, 1) and 2) are in the project area directory. I modified 2) and 4) to contain the IP address of my machine (see previous post). I have not modified 1) and 3), and in fact they do not specifiy node details.

BTW, the target.conf file is not part of the SCNodeI0.tgz, i.e. it is not copied to /root/sisp.

>> Also knowing the versions of kernel,gcc,glibc,pthread library on ur machine would help alot.

The kernel:

[root@mymachine bin]# uname -a
Linux mymachine 2.6.9-42.EL #1 Wed Jul 12 23:16:43 EDT 2006 i686 i686 i386 GNU/Linux

gcc:

[root@mymachine ~]# rpm -q gcc
gcc-3.4.5-2

glibc:

[root@mymachine ~]# rpm -q glibc
glibc-2.3.4-2.19

pthread:

[root@mymachine ~]# getconf GNU_LIBPTHREAD_VERSION
NPTL 2.3.4

>> BTW right now ASP supports RHEL4. Is the version of linux you are using different?

I have RHEL 4.0.3.

BR

Jozsef

Submitted by harikrishna_gp on Thu, 2006-11-02 16:59.

Does your network has any firewall enabled?
If yes can you run 'iptables -F' command as root in your machine.
GMS needs multicasting support and b ecause the firewall is enabled it fails.

HTH

Submitted by jozsefbiro on Mon, 2006-11-06 07:06.

Hi,

Yes, that solved the problem, thanks!

And this is again a reminder to myself that I should always carefully read messages, as you have already guessed this correctly in your first reply, but I just ignored it, can't remember why (probably because I thought that it was irrelevant in a single node configuration...) Anyway, thanks again, things seems to run smoothly now.

BR József

Submitted by harikrishna_gp on Mon, 2006-11-06 16:46.

Thanks for letting me know that it is working for you.

Happy evaluation,
Hari