Lot of kernel errors after setup iscsi storage

[smutel]

Hello,

Since I setup some iscsi storage in Proxmox Version 6, I have a lot of error in my server logs.

In dmesg or /var/log/messages or /var/log/kern.log, I have a lot of:

[2572603.021727] sd 10:0:0:2: [sdn] Unit Not Ready
[2572603.038628] sd 10:0:0:2: [sdn] Sense Key : Illegal Request [current]
[2572603.047133] sd 10:0:0:2: [sdn] Add. Sense: Logical unit not supported
[2572603.055823] sd 10:0:0:2: alua: rtpg failed
[2572603.056208] sd 10:0:0:2: [sdn] Read Capacity(16) failed: Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
[2572603.063090] sd 10:0:0:2: [alua] Sense Key : Illegal Request [current]
[2572603.063100] sd 10:0:0:2: [alua] Add. Sense: Logical unit not supported
[2572603.092962] sd 10:0:0:2: [sdn] Sense Key : Illegal Request [current]
[2572603.101739] sd 10:0:0:2: [sdn] Add. Sense: Logical unit not supported
[2572603.110618] sd 10:0:0:2: [sdn] Read Capacity(10) failed: Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
[2572603.122342] sd 10:0:0:2: [sdn] Sense Key : Illegal Request [current]
[2572603.131422] sd 10:0:0:2: [sdn] Add. Sense: Logical unit not supported
[2572603.143115] sd 10:0:0:254: alua: supports implicit TPGS
[2572603.150237] sd 10:0:0:254: alua: device naa.624a93705c62ab4d4df24c8c00028ec6 port group 1 rel port 20
[2572603.176950] sd 10:0:0:254: alua: port group 01 state A non-preferred supports tolUSNA

I can add some configuration file to this thread if needed.

Thanks.

 

[bbgeek17]

Your iSCSI storage is not working properly. The disk/LUN exposed from your storage system is not ready.

---

Blockbridge: Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[smutel]

Your iSCSI storage is not working properly. The disk/LUN exposed from your storage system is not ready.

---

Blockbridge: Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

Thank you for your answer but it will not help me to resolve what's wrong.

 

[bbgeek17]

Thank you for your answer but it will not help me to resolve what's wrong.

The answer is as expansive as the information you provided about your problem.

These messages are actual responses from whatever storage system you are using. It literally sent back "Unit Not Ready" and others.
[2572603.021727] sd 10:0:0:2: [sdn] Unit Not Ready
[2572603.038628] sd 10:0:0:2: [sdn] Sense Key : Illegal Request [current]
[2572603.047133] sd 10:0:0:2: [sdn] Add. Sense: Logical unit not supported

Based on the information you provided : it is not a network issue, not an iSCSI issue, its a SCSI layer response. It means you should be looking at your mystery storage logs and/or opening a case with your storage vendor.

---

Blockbridge: Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[smutel]

Thank you. I will investigate what's wrong on my mystery storage.

 

[smutel]

I am still investigating why I have these messages.

In another server I have only these lines:

Apr 25 16:47:39 kernel: [2864560.787305] sd 8:0:0:1: alua: supports implicit TPGS
Apr 25 16:47:39 kernel: [2864560.792895] sd 8:0:0:1: alua: device naa.624a93705c62ab4d4df24c8c000150f3 port group 1 rel port 20
Apr 25 16:47:39 kernel: [2864560.876129] sd 8:0:0:1: alua: port group 01 state A non-preferred supports tolUSNA

Is-it ok ?

These lines are generated by the iscsiadm rescan command which is executed every 10 seconds (https://github.com/proxmox/pve-stor...103bb9ce66fed/PVE/Storage/ISCSIPlugin.pm#L136).
Why this command is executing every 10 seconds ?

Thanks.

 

[bbgeek17]

The kernel messages you posted are informational, nothing to worry about. If you search for "alua TPGS" the first google result should be a nice vmware article explaining the terms.
The iscsiadm rescan is there to pickup any changes that were initiated from storage side, ie LUN resize. Its run in a loop to hide the complexity of managing storage from most people who dont need to deep dive into it.

---

Blockbridge: Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[smutel]

I understood that these logs are informational but how to handle 3 lines of logs per iscsi session every 10 seconds without fulfill the log partition ? Having a big log partition ? Specific rsyslog config ? Specific logrotate config ?

 

[bbgeek17]

If you are not satisfied with how PVE iSCSI plugin handles automatic scanning you can configure iSCSI directly via Linux native tools, ie iscsiadm, and just point your LVM/ZFS/etc to that storage.

In a standard installation there is no log partition, the logs are collocated on your root partition. The logs are already automatically rotated and compressed, with old ones being aged out. You can certainly research how to make the rotation more aggressive and/or send certain messages to another log file thats rotated more often. This would be done via native Debian/Linux administration methods. Be careful to properly enter such customization to avoid it being overwritten.

Some back of the napkin math 322 characters in those 3 lines, every 10 seconds:
322*6(per min)*60(per hour)*24(per day)*7(week)=19474560(~20mb) which will be rotated and compressed. Probably not your biggest disk space hog.

---

Blockbridge: Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox