iSCSI errors on storage server

krpa

Active Member
May 10, 2013
1
0
41
Hi

I have Debian Wheezy based iSCSI storage server running LIO iSCSI target and configured with targetcli.
Also I have two Proxmox VE 2.3 nodes in non-HA cluster configuration. Nodes are up to date.

When I add iSCSI target to cluster I have bunch of rx_data iSCSI errors (see below)

Target works fine (LVM on top) but those errors on storage server are annoying and produce lots of noise in syslog

I have traced the problem to /usr/bin/pvestatd, /usr/share/perl5/PVE/Storage/ISCSIPlugin.pm and iscsi_test_portal function which connects to iSCSI portal and disconnects but LIO target expects full ISCSI_HDR_LEN (drivers/target/iscsi/iscsi_target_login.c in kernel tree)

Quick fix is to return 1; from iscsi_test_portal instead of return $p->ping($server); but I am not sure will any other functionality suffer if we assume that target is always connectable.

Igor


/var/log/syslog on storage sever:
May 10 11:22:21 storage kernel: [609526.746899] rx_data() returned an error.
May 10 11:22:21 storage kernel: [609526.746995] iSCSI Login negotiation failed.
May 10 11:22:31 storage kernel: [609535.900565] rx_data() returned an error.
May 10 11:22:31 storage kernel: [609535.900591] iSCSI Login negotiation failed.
May 10 11:22:41 storage kernel: [609546.016885] rx_data() returned an error.
May 10 11:22:41 storage kernel: [609546.016912] iSCSI Login negotiation failed.
 
Hi,

same problem here with Proxmox 3.0 RC2 from today

Code:
[49689.558598] iSCSI: Login negotiation failed from [192.168.3.2]
[49699.792302] rx_data() returned an error.
[49699.796311] iSCSI: Login negotiation failed from [192.168.3.2]
[49709.022337] rx_data() returned an error.
[49709.026330] iSCSI: Login negotiation failed from [192.168.3.2]
[49719.258469] rx_data() returned an error.
[49719.262418] iSCSI: Login negotiation failed from [192.168.3.2]
[49729.487455] rx_data() returned an error.
[49729.491443] iSCSI: Login negotiation failed from [192.168.3.2]

Proxmox has 192.168.3.2
 
but LIO target expects full ISCSI_HDR_LEN (drivers/target/iscsi/iscsi_target_login.c in kernel tree)

Maybe you can fix the kernel code:

Code:
    if ([URL="http://lxr.free-electrons.com/ident?i=rx_data"]rx_data[/URL]([URL="http://lxr.free-electrons.com/ident?i=conn"]conn[/URL], &iov, 1, [URL="http://lxr.free-electrons.com/ident?i=ISCSI_HDR_LEN"]ISCSI_HDR_LEN[/URL]) <= 0) {

I guess it is not necessary to log an error if the connection is immediately closed (I guess rxdata returns 0 in that case)?
 
Any update on this front? I have a ZFS on Linux box and I am attempting to set up an iSCSI target for Proxmox. Has anyone found a fix for this and/or does this even affect performance? It definitely clutters the syslog.
Does anyone use a different iSCSI target?

Thanks!
 

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!