====== iSCSI Enterprise Target ======

===== Konfiguration =====
==== Target ====
=== LUN ===
  * Immer ScsiSN und ScsiId setzen (evtl. auf die ersten 16 Zeichen der LV UUID)
  * Eventuell Type=blockio verwenden

  In blockio mode, it defines a mapping between a "Logical Unit Number" <lun> and a given block device <device>. 
  This mode will perform direct block i/o with  the device,  bypassing  page-cache for all operations. 
  This allows for efficient handling of non-aligned sector transfers (virtualized environments) and large block transfers (media servers). 
  This mode works ideally with high-end storage HBAs and for applications that either do not need caching between application and  disk or need the large block throughput.

  * Eventuell eine Kombination aus Type=fileio und IOMode=wb testen

  Setting IOMode to "wb" will enable write-back caching.
  Please note that this could lead to data loss e.g. in case of power outages,
  so it is strongly recommended to use an UPS when enabling write-back caching.

===== Tuning =====
==== Scheduler ====
Immer **deadline** verwenden
<code bash>
cat /sys/block/<device>/queue/scheduler
echo deadline > /sys/block/<device>/queue/scheduler
</code>

==== Netzwerkkarte ====
=== e1000 ===
Verringern der Latenz zur Steigerung des Durchsatzes:
  * Disable TOE if enabled, you can also try disabling TSO as well.
  * options e1000 InterruptThrottleRate=1,1

===== Betrieb =====
==== Vergrößern einer LUN ====
  - Unmount the volume on the initiator, if needed.
  - Logout and delete the node on the client initiator, so it's no longer accessed.
  - On your fileserver, delete the LUN from the target using ietadm (ietadm --op delete --tid=X --lun=Y). You don't need to delete the whole TID, just the affected LUN. You get this information from /proc/net/iet/volume.
  - Expand/modify/whatever the underlying volume on the server.
  - Dynamically add the LUN back: ietadm --op new --tid=X --lun=Y --params Type=fileio,Path=/dev/foo/bar
  - On the client initiator, add the node again and log back in.
  - You may then need to resize the volume on the client with resize2fs or something similar.