"Wait For RAM Disk Response" while capturing a Linux image

[Laurent MARTIN (TS Fran]

Hello,

I need some help with Insight Control for Linux.
I'm using the last version (6.2) of the software with updated MIBs and PSP on managed servers (RHEL 5.4 on Blade G6 with 2 mezzanine cards).
Everything is working fine, except when i'm trying to capture a Linux image.
The "Wait For RAM Disk Response" step fails.
I'm not using PXE boot method as i'm not using DHCP, I'm using Virtual Media method.
I configured the network stuffs.
Everything seems to be working fine : the captured server is restarting normally (using Virtual Media), but the "Wait For RAM Disk Response" still fails.

The is no enabled firewall. The MP association in the Network menu is correct. I download the last initrd image.

I have no idea what's the problem.

[Rigoberto Corujo]

Hello,

When you observe the console as the system is booting, you should see a message that says "Welcome to IC-Linux Virtual Media Deployment Tool" and then a messaging saying that the kernel and initrd.img are loading. After they load, you should see the kernel booting and will get to a point where, if you press "Enter", you should see the busybox shell prompt. Are you getting this far?

Rigoberto

[Laurent MARTIN (TS Fran]

Hello,

In fact, the server isn't restarting, I have a "runaway loop" issue while booting.

Here are the logs of the capturing tasks (BTW "x.x.x.112" is the CMS IP, "x.x.x.105" is the captured server IP, initrd.img has been replaced as mentionned in the documentation (initrd rhel5 x86_64)) :

- STEP 1 : Power Down Server
Checking to see if power is on.
Attempting to power down via management processor
Completed successfully
- STEP 2 : Configure System to Boot from Network
Configuring system to boot from Virtual Media.

Removing Virtual Media ISO
Executing: /bin/bash -c /bin/rm -f /opt/repository/iso/00000017a47710380000000300000073-*.iso
Completed successfully.
Booting to 32-bit ramdisk.

Boot Loader Config File:

PROMPT 1
DISPLAY boot.msg
TIMEOUT 100
DEFAULT default
LABEL default
kernel kernel
append initrd=initrd.img,0000000300000073.img root=/dev/ram0 rw mem=60G ramdisk_size=130000 boot_method=VirtualMedia SSHD=y SCRIPTNAME=toolkit teid=00000017a47710380000000300000073 IMAGESERVER=x.x.x.112 BOOTIF=00-17-a4-77-10-22 IPADDR=x.x.x.105 NETMASK=255.255.255.0 HOSTNAME=myhostname105 GATEWAY=x.x.x.5 NAMESERVERS= DOMAINS=
LABEL maintenance
kernel kernel
append initrd=initrd.img,0000000300000073.img root=/dev/ram0 teid=0 rw mem=60G ramdisk_size=130000 boot_method=VirtualMedia SSHD=y SCRIPTNAME=toolkit IMAGESERVER=x.x.x.112 BOOTIF=00-17-a4-77-10-22 IPADDR=x.x.x.105 NETMASK=255.255.255.0 HOSTNAME=myhostname105 GATEWAY=x.x.x.5 NAMESERVERS= DOMAINS=
LABEL local
localboot 0


Creating Virtual Media ISO.
Executing: /opt/hptc/bin/create_iso --iso=/opt/repository/iso/00000017a47710380000000300000073-RWlBve8l3HDCec67.iso --loader=/opt/repository/boot/virtualmedia/isolinux.bin --config=/opt/repository/boot/virtualmedia/00000017a47710380000000300000073 --kernel=/opt/repository/boot/IcleToolkitBoot/kernel --initrd=/opt/repository/boot/IcleToolkitBoot/initrd.img,/opt/repository/boot//IcleToolkitBoot//ramdisk/0000000300000073.img --bootmsg=/opt/repository/boot/virtualmedia/boot.msg
Created /opt/repository/iso/00000017a47710380000000300000073-RWlBve8l3HDCec67.iso
Completed successfully.
Booting to 32-bit ramdisk.
- STEP 3 : Set System to One Time Network Boot
Ejecting Virtual Media CD
Management Processor is iLO2
Inserting CD using https ...
Successfully inserted Virtual Media CD
Setting one time CD-ROM boot.
Completed successfully.
- STEP 4 : Power On Server
Checking to see if power is on.
Completed successfully
- STEP 5 : Wait for RAM Disk Response
Waiting for RAM disk to respond.
Server did not respond in allotted time.
An installation failure occurred. The following provide the most likely reasons for the failure:

* Did you select the correct boot MAC address in the Options -> IC-Linux -> Network Configuration Editor task?
* Did you specify the correct IP address in the Options -> IC-Linux -> Network Configuration Editor task?
* Did you specify the correct net mask in the Options -> IC-Linux -> Define Networks task?
* Is the Ethernet cable firmly connected to the correct NIC?
* If the CMS and target server are on different subnets, did you specify the correct gateway in the Options -> IC-Linux -> Define Networks task?
* If the CMS and target server are on different subnets, is the correct gateway is specified in the iLO? To verify, connect to the iLO with a browser and select Administration->Network.
* Is the port specified by REPOSITORY_HTTPS_PORT in /opt/mx/icle/icle.properties open in the CMS.s firewall? The default value for this port is 60002.
* Does the server require a special initrd file to support the NIC or disk device? See the Insight Control for Linux Support Matrix to verify.
* Is the server's iLO on a different network than the CMS's management NIC? Use the VIRTUAL_MEDIA_REPOSITORY_IP variable in /opt/mx/icle/icle.properties to specify an alternate IP address to reach the CMS.

[Rigoberto Corujo]

Hi Laurent,

At this point, you would be booting to the IC-Linux RAM disk to do the capture, not the Red Hat RAM disk, so the Red Hat rhel5 x86_64 initrd is not the issue.

So you are saying that if you open up a Remote Console that you never see the IC-Linux RAM disk boot?

Can you please try resetting the iLO using the following steps?

1. Log on to the iLO.

ssh -l

2. At the iLO prompt, type:

reset map1

After you have reset the iLO wait about 30 seconds and try the capture again.

Rigoberto

[Laurent MARTIN (TS Fran]

Hello,

I did what you've told me (iLO reset), but it doesn't change anything : same problem.

While capturing and waiting for RAM disk response, I checked the integrated remote console of the iLO of the captured server, and I noticed I have a runnaway loop (See attached file).

[Rigoberto Corujo]

Is the firmware on the server up-to-date?

[Laurent MARTIN (TS Fran]

Yes ! I alreaday checked that

[Donna Firkser]

Laurent,

We've notified L3 support of this issue and they have requested that you create a case with HP so that this issue can be tracked properly.

In the interim, L3 is reviewing the information and will work with us (the IC-Linux development team) to help resolve this issue.

Thanks,
Donna

[Donna Firkser]

Laurent,

What's the exact model number of your G6 Blade? If we have the exact same server, I'm going to see if we can reproduce this issue.

Thanks,
Donna

[Laurent MARTIN (TS Fran]

Hello,

ProLiant BL-460c G6
Part Number 507864-B21

[Laurent MARTIN (TS Fran]

BTW, do you have any phone number or email@ where i can request a L3 support ?

Thanks,

Laurent

[Donna Firkser]

Thanks for the server information. Let me talk to support to see if I can find a number for you to call.

What company do you work for?

Thanks,
Donna

[Donna Firkser]

Hi Laurent,

Can you also let us know what type of mezzanine cards are in your BL460c G6 blade server. You should be able to get this information from the OA.

Thanks,
Donna

[Laurent MARTIN (TS Fran]

Hello,

I work for TS (HP France).
Here are the mezzanine cards :
406771-B21 - HP NC326m PCI Express Dual Port 1Gb Server Adapter.

Thanks for your help.

[Mitchell Kulberg]

Hi Laurent,

Going back a few posts. Can you explain what you mean by a "runaway loop" problem?

Your job is failing because the server is not properly booting up into the IC-Linux ramdisk. The CMS is waiting for teh ramdisk to come up and contact it, but that is never happening.

Please explain more abuot exactly what is happening when your server reboots.

You should see the server shut down, then power back on, run power on self test, and then boot the IC-Linux ramdisk via virtual media.

Please describe what yuo are seeing.

Thanks,
Mitch

[Laurent MARTIN (TS Fran]

Hi Mitchell,

When running the Linux image capture, the task fails at "Wait for RAM Disk Response" step.
I see the server shut down, then power back on, run power on self test, and then boot the IC-Linux ramdisk via virtual media. But when booting the IC-Linux ramdisk I can see "request_module: runnaway loop modprobe binfmt-464c" (See the attached screenshot).

[William Athanasiou]

Looks like the kernel is failing to boot successfully very early in the startup process. Were you able to run a bare metal discovery process on this node? Could you try deleting the node from SIM and then selecting the Management processor and running an "Initiate bare metal discovery" on it. The kernel used for discovery and capture is the same 32bit kernel. I'm surprised one worked, and the other didn't.

[Laurent MARTIN (TS Fran]

In fact I modified the initrd.img as advised in the documentation for Blades G6.
I don't know if the initrd.img is failing, but I checked that the image is the good one, and it seemed so.

[Donna Firkser]

Laurent,

A few more questions for you.

1. Was this a fresh install of IC-Linux 6.2 or an upgrade?

2. Prior to doing the IC-Linux capture, did you use IC-Linux "6.2" to install RHEL5.4 on the blade server? Or was this a pre-installed server that you're trying to capture?

3. Did you try deleting the server from HP SIM and doing another bare metal discovery (as Will Athanasiou suggested)?

Thanks,
Donna

[Laurent MARTIN (TS Fran]

Hello,

Here are my answers :

1. It's a fresh install of IC-Linux 6.2 (a previous version was installed but I uninstalled it first). The capture never worked.

2. All servers were pre installed. IC-Linux purpose is to monitor these servers, but now we need also the capture function in order to be able to "backup" the servers.

3. I didn't try deleting the server in HP SIM because I'm not onsite at the moment. I need a few weeks more to get back there.

Thanks,

[Mitchell Kulberg]

Laurent,

To to add another data collection point to this, could you please try the "Boot to Linux rescue mode" function on this server and see if it exhibits the same behavior?

Thanks,
Mitch

[Laurent MARTIN (TS Fran]

Hello, i tried everything, but the server successfully boots when I'm not trying to capture it...

[Laurent MARTIN (TS Fran]

I also changed the initrd of Insight Control Repository by this one : ftp://www.hp.com/pub/softlib2/software1/pubsw-linux/p2057226266/v62035/initrd-rh5u4-x86_64.img

[Mitchell Kulberg]

OK, need some more clarification.

First - when you say it boots ok unless you're capturing. That means you are able to "Boot to Linux rescue mode", right? Can you see the server's filesystem from there? If you go to the console, you should see a prompt that says something like "Press ENTER to continue using this console". Press enter and you'll get a prompt. Then, under that ramdisk filesystem I think you'll see the local files on the server under /target or /TARGET or something like that (I'm not in office now to check). Maybe just do a 'mount' command to see where it's mounted. If it's not mounted, try to mount them to verify that you can indeed read the filesystem on the server from the HP ramdisk.

ALSO - When you said you downloaded a new initrd, can you explain more? Those are for OS installations only and are not related to capture at all. I just want to make sure you didn't put it somewhere it didn't belong.

Thanks,

[Laurent MARTIN (TS Fran]

Hello,

I mean, I succedeed in booting linux in rescue mode !
No problem at all.

I downloaded a new initrd because it's described in the User Guide.
I replaced the /opt/repository/boot/IcleToolkitBoot/initrd.img with the one I downloaded.
BTW, it didn't worked either with the original initrd.img

Thanks.