all lists on lists.proxmox.com
 help / color / mirror / Atom feed
From: "Max R. Carrara" <m.carrara@proxmox.com>
To: pve-devel@lists.proxmox.com
Subject: [PATCH storage 0/9] Fix ZFS-over-iSCSI plugin w/ LIO Provider Issues
Date: Thu, 25 Jun 2026 18:40:52 +0200	[thread overview]
Message-ID: <20260625164110.706694-1-m.carrara@proxmox.com> (raw)

Fix ZFS-over-iSCSI plugin w/ LIO Provider Issues
================================================

When giving the ZFS-over-iSCSI plugin with the LIO iSCSI provider a spin
on our test suite, th caching logic in the LIO module causes certain
LUNs / ZFS zvols to end up in an inconsistent state -- sometimes.

Things I have noticed going wrong are mainly dangling LUNs / zvols --
volumes that were not deleted correctly -- and the plugin sometimes
preventing me from creating a new disk for a VM where I just recently
deleted a disk, because it thinks the deleted one still exists.

The latter scenario fixes itself once the cache is invalidated after 15
seconds, but the former case requires manual fixing on the remote host
using `targetcli`.

Instead of trying to fix the cache invalidation logic, this series
removes the cache altogether, resulting in the plugin running stable
even when a lot of operations are being performed.

Another issue that may occur is not being able to delete a volume with a
"nonstandard" name, e.g.:

    pvesh create /nodes/localhost/storage/zfs-iscsi-lio/content \
        --filename vm-110-disk-2-foo-bar-baz.raw \
	--size 4G \
	--vmid 110

Note that the main culprit here is actually the '.raw' suffix.

For ZFS-over-iSCSI, all volumes and LUNs are usually created in the
format of 'vm-100-disk-0', 'vm-420-disk-2', etc. The respective parsing
logic is carefully relaxed in this case, allowing such volumes to be
deleted again.

All of this is addressed in the first three patches. The remaining
patches modernize / clean up the code of the LIO module to bring it in
line with our contemporary code.

This entire series was tested using our test suite while developing it,
but some additional testing wouldn't hurt.

Summary of Changes
------------------


pve-storage:

Max R. Carrara (9):
  luncmd: lio: fix various errors by removing caching mechanism
  luncmd: lio: fix LUN ops failing when passing nonstandard LUN path
  luncmd: lio: remove rest of old parsing logic
  luncmd: lio: modernize helpers
  luncmd: lio: modernize & clean up `list_view()` and `list_lun()` subs
  luncmd: lio: modernize sub `create_lun()`
  luncmd: lio: modernize sub `delete_lun()`
  luncmd: lio: modernize remaining LUN command subroutines
  luncmd: lio: use v5.36 and use subroutine signatures

 src/PVE/Storage/LunCmd/LIO.pm | 521 ++++++++++++++++------------------
 1 file changed, 243 insertions(+), 278 deletions(-)


Summary over all repositories:
  1 files changed, 243 insertions(+), 278 deletions(-)

--
Generated by murpp 0.11.0




             reply	other threads:[~2026-06-25 16:41 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-06-25 16:40 Max R. Carrara [this message]
2026-06-25 16:40 ` [PATCH pve-storage 1/9] luncmd: lio: fix various errors by removing caching mechanism Max R. Carrara
2026-06-25 16:40 ` [PATCH pve-storage 2/9] luncmd: lio: fix LUN ops failing when passing nonstandard LUN path Max R. Carrara
2026-06-25 16:40 ` [PATCH pve-storage 3/9] luncmd: lio: remove rest of old parsing logic Max R. Carrara
2026-06-25 16:40 ` [PATCH pve-storage 4/9] luncmd: lio: modernize helpers Max R. Carrara
2026-06-25 16:40 ` [PATCH pve-storage 5/9] luncmd: lio: modernize & clean up `list_view()` and `list_lun()` subs Max R. Carrara
2026-06-25 16:40 ` [PATCH pve-storage 6/9] luncmd: lio: modernize sub `create_lun()` Max R. Carrara
2026-06-25 16:40 ` [PATCH pve-storage 7/9] luncmd: lio: modernize sub `delete_lun()` Max R. Carrara
2026-06-25 16:41 ` [PATCH pve-storage 8/9] luncmd: lio: modernize remaining LUN command subroutines Max R. Carrara
2026-06-25 16:41 ` [PATCH pve-storage 9/9] luncmd: lio: use v5.36 and use subroutine signatures Max R. Carrara

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260625164110.706694-1-m.carrara@proxmox.com \
    --to=m.carrara@proxmox.com \
    --cc=pve-devel@lists.proxmox.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.
Service provided by Proxmox Server Solutions GmbH | Privacy | Legal