Question

HP DL360p Gen8 Disks not Visible?


Badge +4
I have CE on an HP DL360 Gen8. My CE installation is not nested. Neither the SSD nor the HDD are detected. Both are in their own RAID0 with logical disk. I read elsewhere that this hardware works. Am I missing something?

This topic has been closed for comments

26 replies

Userlevel 7
Badge +25
Betting your scsi card isn't being detected and the vols aren't being presented to install. If you login as root (not install) and fdisk what do you see?Guessing nothing which means you would need to get a kernel mod for that scsi adapter. Attaching the disks via ahci an option?
Userlevel 6
Badge +29
have u run the latest service pack for proliant firmware updater?
Badge +4
I can not attach the disks directly to the SATA controller, they only go throught the RAID controller. Looks like the disks are present under fdisk. So, if they're here, why don't I see them for creating a storage pool?

/dev/sda is the 200GB SSD
/dev/sdb is the 500GB HDD
/dev/sdc is the 128GB boot USB

Badge +4
I have not run the latest HP updaters. I suppose I could, but it seems that the disks are visible. Do you think it would still make a difference?
Userlevel 7
Badge +25
what does firstboot.out say?
Badge +4
The output is too large to post here, so here's a link to the file. https://drive.google.com/file/d/0BxHrwU_BcW5Wak9mTmNpRllBQ28/view?usp=sharing
Userlevel 7
Badge +25
Sorry blacklisted for me where I am for now so working blind. Anything useful you can see in disk detection? errors or failures?
Badge +4
The only error I see is the following many times:

INFO: Cmd [['/usr/bin/ssh -i /root/firstboot/ssh_keys/nutanix -o StrictHostKeyChecking=no -o NumberOfPasswordPrompts=0 -o UserKnownHostsFile=/dev/null nutanix@192.168.5.2 "ls /tmp/svm_boot_succeeded"']] failed 9 times with reason [Warning: Permanently added '192.168.5.2' (RSA) to the list of known hosts.Nutanix Controller VMls: cannot access /tmp/svm_boot_succeeded: No such file or directory]. Will retry in 5 seconds
Userlevel 7
Badge +25
Thats it trying to start the CVM at the end of the install. You will want to look way earlier around detecting disks. maybe look for sda or something.
Badge +4
Minus what I already posted:
INFO: Running cmd ['grep "release 7" /etc/redhat-release']INFO: Running cmd ['service ntpd stop']INFO: first_boot_config = {"svm_default_gw": "192.168.2.1","installer_path": "/mnt/local/images/svm/2016.03.13-stable/nutanix_installer_package.tar","dell_hyp_lun_label": null,"phoenix_version": "phoenix-e0ced8f953d2eb629dbffbc56a9289cbd4e6ead5_e0ced8f9","svm_ip": "192.168.2.191","hyperv_binaries": null,"dns_ip": "8.8.8.8","svm_version": "2016.03.13-stable","fio_detected": false,"cluster_id": 3302031626616800440,"svm_num_vcpus": 8,"partition_table_type": null,"vibs": [],"foundation_ip": null,"node_name": "e3ff3b97-A","boot_disk_model": "Ultra Fit","hyp_type": "kvm","dell_factory_lun_label": null,"foundation_version": "unknown","boot_disk_sz_sectors": null,"use_ten_gig_only": false,"ce_eula_viewed": true,"boot_disk_info": null,"use_vmdk_svm_disk": false,"boot_disk_sz_GB": 124.21850726400001,"cluster_name": "NTNX","model_string": "CommunityEdition","diagnostics_path": "/mnt/local/images/diagnostics/kvm","dell_logfile_error": null,"use_vmfs_datastore": false,"hyp_install_type": "clean","factory_iso_path": null,"dell_phoenix_lun_label": null,"block_id": "e3ff3b97","ptable_end_sector": null,"node_position": "A","bootbank": null,"hyp_image_path": "/mnt/local/images/hypervisor/kvm/20150120/diamond_image/kvm.img.xz","dell_hyp_lun_index": null,"host_ip": "192.168.2.190","hypervisor_hostname": null,"network_devices": {"eno1": {"vendor_id": "14e4","mac_addr": "2c:44:fd:99:84:b8","device_id": "168e","vendor_string": "Broadcom"},"eno2": {"vendor_id": "14e4","mac_addr": "2c:44:fd:99:84:bc","device_id": "168e","vendor_string": "Broadcom"}},"hyp_version": "20150120","dell_error_file": null,"boot_disk": "sda","create_1node_cluster": true,"svm_subnet_mask": "255.255.255.0","ce_eula_accepted": true,"host_subnet_mask": "255.255.255.0","dell_hyp_image": null,"node_serial": "c813a4cc-66db-45dd-9649-bed986affc50","storage_passthru": true,"node_uuid": "6886ca73-7ba9-45aa-9d3d-b3083d33b36a","monitoring_url_root": null,"dell_logfile_info": null,"dell_factory_run_level": 0,"dell_phoenix_lun_index": null,"dell_perc_boot_device": true,"dell_factory_lun_index": null,"default_gw": "192.168.2.1","dell_perc_adapter": null,"model": "USE_LAYOUT","svm_gb_ram": 16,"svm_install_type": "clean"}INFO: Running cmd ['restorecon -R /root/.ssh']INFO: Running cmd ['ovs-vsctl add-br br0']INFO: Running cmd ['ethtool eno1']INFO: Running cmd ['ethtool eno2']INFO: Creating ifcfg script for interface 'eno2'INFO: Running cmd ['/sbin/ifup eno2']INFO: Creating ifcfg script for interface 'eno1'INFO: Running cmd ['/sbin/ifup eno1']INFO: Running cmd ['restorecon -R /etc/sysconfig/network-scripts']INFO: Running cmd ['ovs-vsctl add-bond br0 bond0 eno2 eno1']INFO: Running cmd ['/sbin/ifdown br0']INFO: Running cmd ['/sbin/ifup br0']INFO: Expanding boot partition. This may take some time.INFO: Running cmd ['/usr/bin/nohup /sbin/resize2fs /dev/sda1 &>/dev/null &']INFO: Running cmd ['/sbin/iscsi-iname']INFO: Running cmd ['restorecon -R /etc/iscsi']INFO: Running cmd ['virsh net-destroy default']INFO: Running cmd ['virsh net-undefine default']INFO: Running cmd ['virsh net-destroy VM-Network']INFO: Running cmd ['virsh net-undefine VM-Network']INFO: Running cmd ['virsh net-define /root/net-VM-Network.xml']INFO: Running cmd ['virsh net-start VM-Network']INFO: Running cmd ['virsh net-autostart VM-Network']INFO: Running cmd ['virsh net-destroy NTNX-Local-Network']INFO: Running cmd ['virsh net-undefine NTNX-Local-Network']INFO: Running cmd ['virsh net-define /root/net-NTNX-Local-Network.xml']INFO: Running cmd ['virsh net-start NTNX-Local-Network']INFO: Running cmd ['virsh net-autostart NTNX-Local-Network']INFO: Running cmd ['echo "$(hostname)-CVM"']INFO: Running cmd ['sed -i "s/NTNX-CVM/NTNX-e3ff3b97-A-CVM/" /root/NTNX-CVM.xml']INFO: Running cmd ['sed -i "//d" /root/NTNX-CVM.xml']INFO: Running cmd ['virsh define /root/NTNX-CVM.xml']INFO: Running cmd ['virsh start "NTNX-e3ff3b97-A-CVM"']INFO: Running cmd ['virsh autostart "NTNX-e3ff3b97-A-CVM"']INFO: Run ssh cmd on SVM
INFO: Run ssh cmd on SVMINFO: Running cmd ['/usr/bin/ssh -i /root/firstboot/ssh_keys/nutanix -o StrictHostKeyChecking=no -o NumberOfPasswordPrompts=0 -o UserKnownHostsFile=/dev/null nutanix@192.168.5.2 "ALL_IPS=\`hostname -I\`; CVM_IP=\`for i in \$ALL_IPS; do echo \$i | grep -v 192.168.5 | grep -v 127.0.0.1; done\`; if [ "\$CVM_IP" == "" ]; then echo -n "Could not detect CVM IP."; exit 1; fi; source /etc/profile.d/nutanix_env.sh; echo Y | cluster -s \$CVM_IP create; sleep 30; source /etc/profile.d/nutanix_env.sh; sleep 5; ncli cluster add-to-name-servers servers=8.8.8.8"']INFO: Cluster created and DNS server (8.8.8.8) added to cluster settings.INFO: Running cmd ['touch /root/.firstboot_success']INFO: Running cmd ['service ntpd start']
Badge +4
I updated the HP firmware. It made no difference.
Userlevel 6
Badge +29
  - thoughts?

 -Thanks for updating the firmware. HP is notorious for having all sort of "firmware fun", so running SPP/PSP to the latest cookbook fixes a lot of drama a lot of the time.

Next up, Do you have iLO access, or just direct monitor access? Can you post some screenshots from your HBA configuration?

Jon
Userlevel 3
Badge +17
If there are disk detection issues, we should see something in the following:

On host, we can see the CVM's serial output: /tmp/NTNX.serial.out.0 Bottom of the file should look similar to:
...Starting httpd: [ OK ]Starting crond: [ OK ]nutanix installed:1.1 desired:1.1SELinux policy nutanix 1.1 is already installedStarting HadesHades started on pids [1890, 1903, 1921, 1922]Waiting for Hades initialization to completeStarting GenesisGenesis started on pids [2104, 2116, 2138, 2139]Presuming the CVM is at least booted, check on CVM: ~/data/logs/hades.out - Check for disk visibility/configuration issues ~/data/logs/genesis* - Check for errors regarding disk detection/hades errors
Badge +4
I was able to make some progress. I had this problem, specifically #4 and #5. Because HP, both of my disks ended up with the same serial numbers, and then the HDD appeared as an SSD. I worked through the instructions in that post and was able to get almost all the way.

Now I can make a pool and container, but in Prism, the drive capacity is much lower than expected. I have a 200GB SSD and 500GB HDD, but they show up as 60GB and 250GB. Here's what I get:



BTW, I do have remote ILO access and can screenshot anything needed. 🙂
Userlevel 6
Badge +29
Fantastic progress.

RE 60GB SSD - thats roughly expected, given that the SSD is multi-purpose, 60GB is "persistent storage" in the extent store, but that doesnt account for things like metadata, various caches, and the nutanix OS itself.

check out http://www.nutanixbible.com, there's a diagram in there on how SSD's are partitioned out.

HDD though, thats interesting, as there shouldn't be nearly as much overhead per HDD (just curator takes up some space, but definitely not that much space). You should be able to do something like df -h or list out the partitions to see whats up there.
Badge +4


Ok, I looked at the Nutanix Bible explaination for the SSD use. I agree, it's in the zone. However, the HDD should have most of its space available for the extent store. According to the df-h output, it looks like the 500GB drive (sdb) should read 434GB available.

Um, I noticed a correlation here. Maybe it's just a coincidence. Look at this:


The 434GB I'm expecting minus the 123GB available from sda4 equals 311GB. The 311GB minus the 60GB for the SSD equals 251GB, which is what's being reported for the HDD. Am I crazy? Maybe I've been staring at all of this for too long.
Userlevel 3
Badge +17
No, you're not crazy.

The data disks are not partitioned to account for the space allocated to "meta-data", "oplog", "content cache", "curator", and such.
These are logically calculated within Acropolis. So "df" will show pretty much the drive sizes expected (less formatting, etc.). But, the GUI will show the space allocated for only for the "extent store" (bulk) storage.
Generally, each drive could have such allocations up to 50-100GB for non-extent storage, depending on which bits are allocated on which drive(s).

Does that help?
Badge +4
It does, thank you. It seems the container functions as expected, so I guess I'm done. 🙂
Badge +3
Hi Warthan,

I'm running into exact same original issue as you. No HDD or SSD are being detected by the installer, however fdisk -l clearly shows them there. What did you end up doing to get past that?

I have the same DL360p Gen8 with P420i. I do not, however, have the cache card in mine.
Badge
Hello,

I installed Community edition for POC on 3 DL360 (2x360e 1x360p).

Forgot RAID configuration.

You need to configure your P420i as HBA controller.
https://www.youtube.com/watch?v=JuaezJd4C3I

Best Regards
VBO
Badge +3
Thanks for the link VBO. I did follow these steps and confirmed in SSA that the P420i is now in HBA mode. For some ridiculous reason the installer finds zero disks.
I have two SAS 146GB 10k's and two SATA 120GB SSDs.

Badge +3
I'm now seeing that I did put insufficiently sized drives in, per the minimum req's. So I swapped one of the SSD with a 250GB model. Unfortunately, same result. I figured it would at least list that one as 'matching the minimum requirements'.
Userlevel 5
Badge +9
Hi,
you mention
I have two SAS 146GB 10k's and two SATA 120GB SSDs.

if this is correct, your HDD is also too small, you need > 500GB.
Secondly, what does the /home/install/phx_iso/phoenix/collect_disk_info.py script return?
I'd also suggest you look at other DL360 threads, there are several of them.
Badge +3
Yes I realize I'll need to beef up the HDD capacity as well. However I had hoped to see at least the 250GB SSD listed as a usable candidate, which would at least tell me I've got the controller settings right.

Userlevel 5
Badge +9
Hi,
collect_disk_info.py script seems to list the 250GB SSD at the end, not sure why it is not listed during install, rotational seems to be set to false.
See also https://next.nutanix.com/discussion-forum-14/no-disks-detected-by-installer-28301 for additional tips.
As for the HDD, Nutanix won't work unless it is big enough so you'll need to add it anyway.
I also looked at a list of known issues in this thread https://next.nutanix.com/discussion-forum-14/ce-on-older-hp-g5-g6-g7-servers-6434, but none seems to fit your issue (it is for older HW though).