Summary
From time to time, an agentless job writing to a vStor will perform an incomplete clean of the iSCSI logical units mapped to the proxy from the vStor. This article will describe how to identify such logical units and remove them.
Step By Step
Validate that no agentless backups are running on the enterprise using this proxy. There's little visual difference in a dead LUN and one in active use. It is safer to wait, then risk removing something in use. The Job Monitor can be used to see if there are agentless jobs running (BACKUP_VIRTUALIZATION).
Validate that no iSCSI mappings exist for IA map or restores on this proxy. In the Restore section of the Management Console, right click on the proxy and select Show Mapped Drives. The Job Monitor can be used to identify any suspect Restore jobs.
Login to the proxy and run the command lsscsi. For example:
Krystal:~ # lsscsi
[0:0:0:0] cd/dvd NECVMWar VMware SATA CD00 1.00 /dev/sr0
[30:0:0:0] disk VMware Virtual disk 1.0 /dev/sda
[34:0:0:0] disk LIO-ORG FILEIO 4.0 -
[34:0:0:1] disk LIO-ORG FILEIO 4.0 -
[34:0:0:4] disk LIO-ORG FILEIO 4.0 -
[34:0:0:5] disk LIO-ORG FILEIO 4.0 -
iSCSI devices presented by the vStor will have the vendor ID "LIO-ORG" and model ID "FILEIO". If no such devices are present, there's nothing that needs to be done. If you're comfortable that there are no active iSCSI LUNs on the proxy, run the lsblk command:
Krystal:~ # lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
fd0 2:0 1 4K 0 disk
sda 8:0 0 200G 0 disk
|-sda1 8:1 0 200G 0 part
|-system-swap 254:0 0 2G 0 lvm [SWAP]
|-system-root 254:1 0 40G 0 lvm /
|-system-home 254:2 0 100G 0 lvm /home
sdi 8:128 0 32K 1 disk
|-sdi1 8:129 0 1G 1 part
|-sdi2 8:130 0 152G 1 part
sr0 11:0 1 1024M 0 rom
sda 8:0 0 200G 0 disk
|-sda1 8:1 0 200G 0 part
|-system-swap 254:0 0 2G 0 lvm [SWAP]
|-system-root 254:1 0 40G 0 lvm /
|-system-home 254:2 0 100G 0 lvm /home
sdi 8:128 0 32K 1 disk
|-sdi1 8:129 0 1G 1 part
|-sdi2 8:130 0 152G 1 part
sr0 11:0 1 1024M 0 rom
Dead iSCSI LUNs will have an uncommonly small size, typically 32K, with respect to the size of the partitions listed after it. In the example above the 32K size of sdi is a good indication that this is dead LUN. The much larger size of the partitions after it, sdi1 and sdi2, help confirm that this is a dead LUN. Use the command "lsblk -S" to find the device address for dead LUN:
Krystal:~ # lsblk -S NAME HCTL TYPE VENDOR MODEL REV TRAN sda 30:0:0:0 disk VMware Virtual disk 1.0 spi sdi 34:0:0:0 disk LIO-ORG FILEIO 4.0 iscsi sr0 0:0:0:0 rom NECVMWar VMware SATA CD00 1.00 sata
This listing includes the Vendor and Model ID to help confirm a device is being presented by the vStor. The SCSI address is the value in the HCTL field (HBA, Controller, Target and LUN). The device can be removed with:
echo 'scsi remove-single-device H C T L' > /proc/scsi/scsi
using the values from the HCTL field. For example:
echo 'scsi remove-single-device 34 0 0 0' > /proc/scsi/scsi
Be cautious with this command, since it will also remove active devices if the wrong address is used.
Similar looking devices with appropriate sizes could be LUNs which still have a filebackedlun associated with them on the vStor. Reconfirm that there are no active LUNs due to jobs or IA mappings. On the vStor run the command "vstor volume show" to see if there are any volumes of Type "filebackedlun":
4953 | filebackedlun | 1 | No | 200.00GB | 58.65GB | 141.35GB | /vol/Syncsort//[vmsSyncsort-2]DPXQA@CED10E9F/6000C290-4d58-15a3-5f6a-9df17ea56464.RAW/SSSV_lun_{vmsSyncsort-2}DPXQA_CED10E9F
If a filebackedlun such as this is not associated with any running job, IA mapping or restore, it can be removed with the command "vstor volume delete --id VOLID", using the volume ID from the first column of the listing. In this example:
vstor volume delete --id 4953
Use the earlier section of the article to remove the now dead LUN.