[SOLVED] Proxmox 4.0: Failed to boot after zfs install : cannot import rpool: no such pool

I completely zeroed the disks and reinstalled. Now it works.

Maybe interesting is, that the disk didn't have a partition table on the previous installation with software RAID. It was a software RAID 1 with complete disks (sda and sdb).

But using ZFS shouldn't depending on the previous layout of the disks?
 
3Gb/s is 300MB/s, and since I'm using RAID10 on 4 disks, it could read the data from 2 disks in parallel at the same time. The technical specification of Velociraptor 1TB states that "Data transfer rate (max) Host to/from drive (sustained) - 200 MB/s" so can be 400MB/s. In addition, and it what worries me more, the Intel DC3710 SSD that is acting as cache should exceed the 450MB/s, and having a bottleneck there seems not smart.
Of course it "will work", but I'm wondering why on a "server class" MB they have saved some dollar to use a 3GB/s chip instead of a more performant one.
 
Hi,

I had a similar problem resulting in exactly the same error message as described by the others. I installed PVE 4 on a laptop with one hard disk and one ssd: zraid0 on the only harddisk, which worked perfectly, even reboots and then added the ssd as a caching device (L2ARC and ZIL), which worked until reboot: yields the aforementioned error message. I reinstalled it thrice, no luck (tried also legacy and uefi). I ended up using a very small boot and root LVM (4GB) and ZFS on a big partition with SSD caching. I moved /usr, /var/log and /var/lib/vz on individual ZFS volumes.

Best,
Andreas
 
3Gb/s is 300MB/s, and since I'm using RAID10 on 4 disks, it could read the data from 2 disks in parallel at the same time. The technical specification of Velociraptor 1TB states that "Data transfer rate (max) Host to/from drive (sustained) - 200 MB/s" so can be 400MB/s. In addition, and it what worries me more, the Intel DC3710 SSD that is acting as cache should exceed the 450MB/s, and having a bottleneck there seems not smart.
Of course it "will work", but I'm wondering why on a "server class" MB they have saved some dollar to use a 3GB/s chip instead of a more performant one.

Still, there are 3Gbit per channel, not per controller, right? Regarding DC3710, you will see limits on sequential reads (but they will be off-loaded to main storage if the L2ARC is too busy, I think). Having 300MB/sec of random read I/O is nice enough. You can add 2 of them, smaller size, to keep the costs down.
 
Last edited:
Still, there are 3Gbit per channel, not per controller, right?
Oh, you are right, seems not a big issue then! I'll do some test, if this issue can be solved for me completely erasing the 4 HD. Wondering why now there are 12MB/s sas/sata controllers available. Thanks for the clarification
 
You marked as solved, but today I've tested my hd and have no mdadm raid info, nor it works with a new driver.
a) ZFS Raid0 fails the same
b) ext4 on the same driver works fine
c) erasing (just in case) dd if=/dev/zero of=$YOUR_DEV bs=512 seek=$(( $(blockdev --getsz $YOUR_DEV) - 1024 )) count=1024 did not changed anything
I've the following output of some commands suggested in the thread:
Code:
zpool import
   pool: rpool
     id: 4282105346604124069
  state: ONLINE
 action: The pool can be imported using its name or numeric identifier.
 config:

    rpool       ONLINE
      sda2      ONLINE

#ls /dev/disk/by-id/
ata-HL-DT-ST_DVDRAM_GH24NSC0_K51F2SE4829
ata-WDC_WD1000CHTZ-04JCPV1_WD-WX41E748RVDP
ata-WDC_WD1000CHTZ-04JCPV1_WD-WX41E748RVDP-part1
ata-WDC_WD1000CHTZ-04JCPV1_WD-WX41E748RVDP-part2
ata-WDC_WD1000CHTZ-04JCPV1_WD-WX41E748RVDP-part9
usb-JetFlash_Transcend_16GB_08RGI62F8R9DYG8A-0:0
usb-JetFlash_Transcend_16GB_08RGI62F8R9DYG8A-0:0-part1
wwn-0x5001480000000000
wwn-0x50014ee65a7463e0
wwn-0x50014ee65a7463e0-part1
wwn-0x50014ee65a7463e0-part2
wwn-0x50014ee65a7463e0-part9

# blkid -l -t TYPE="zfs_member"
/dev/sda2: LABEL="rpool" UUID="4282105346604124069" UUID_SUB="6939510755325752309" TYPE="zfs_member" PARTLABEL="zfs" PARTUUID="a5a61a3d-c4dc-4eaa-9b55-cd9cb7cdb359"

# blkid | grep "zfs_member"
/dev/sda2: LABEL="rpool" UUID="4282105346604124069" UUID_SUB="6939510755325752309" TYPE="zfs_member" PARTLABEL="zfs" PARTUUID="a5a61a3d-c4dc-4eaa-9b55-cd9cb7cdb359"

Installation of Proxmox has been done with usb
I'm clueless, frustrated and any help is very appreciated!
(last step is a long dd zero of the hd, just in case... but I'm waiting it to finish)
 
Thanks a lot for the reply, but I thin I don't understand what you mean.
When the boot stops I've the same message as the original poster (except a preceding line, just noticed, with "no controller found" then "loading, please wait" the the output mentioned (I attach a screenshot, don't know how to copy from console to text mode file).
Then if I issue "zpool import" I've the output I've mentioned in my previous post
Code:
# zpool import
   pool: rpool
     id: 4282105346604124069
  state: ONLINE
 action: The pool can be imported using its name or numeric identifier.
 config:

    rpool       ONLINE
      sda2      ONLINE
I've dd zeroed the entire hd but I get the same problem (no no raid-metadata related).
Boot_fail_IMG_20151013_113146_small_low.jpg
 
I've tried after failed boot:
Code:
# zpool import -d /dev/disk/by-id rpool -R /root
# exit
failed to mount root filesystem 'rpool/ROOT/pve-1'
# exit
and it boots! But if I reboot t fails in the same way (busybox)
Any clue of the reason of this problem and how to permanently solve?
Thanks a lot
 
Try this:

Boot the system as you did and then:

Code:
[FONT=courier new]
[/FONT][COLOR=#333333][FONT=Consolas][FONT=courier new]# zpool set cachefile= rpool
# update-initramfs -k `uname -r` -u
# reboot[/FONT]
[/FONT][/COLOR]



 
Yes, a new server Intel S2400SC, iso is the latest 4.0 iso (stable) in a usb stick. If I install with ext4 everything works fine. ZFS raid1 on one disk or 10 on 4 disks has the same problem (haven't tested different ZFS raid, just telling what I tested). Could it be that at boot time is not able to grab some drivers / wrong timing in mounting something? I've a vague memory of some thread with troubles on zfs and something added in a config file to have things go at different moment, but can't find anymore.
 
From your error message I could say that the disk drivers are not available at the zpool import time. What disk controller do you have?

Btw, I've seen in a post above that you have a single WD drive on SATA? How did you do RAID1 or RAID10 then?
 
About my HD configuration, I wrote it wrong, I'm doing RAID0 with one disk, sorry!
If you don't mind, I would open a new thread with similar subject because
a) this has been marked as "solved", so I get less attention
b) my problem is not related to the one had the creator of the thread
Thanks a lot
 
Last edited:

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!