What value, if not persistent?

[ed@zugzwang.com]

Hi,
First let me say that when I have a problem with something, it usually turns out that I am doing something wrong.

I don't understand why the iSCSI initiator is persistant between re-boots for single connections,
but the option is gray for RAID-1. The RAID-1 connection can be re-established with a few mouse clicks so why
can't it be defined to "automount"? The information is still there.

It would seem to be of great value to use a locally attached .img file as the active (for speed) and an iSCSI
disk as the passive (for backup).

But, if the definition is not persistent after re-boot, what's the point?
Am I missing something?
Thanks, Ed

btw- Starwind iSCSIinitiator is far easier to use than the MSinitiator!!

[anton (staff)]

The point was to build experimental feature. We did it but it did not raise wide adoption so we've lost interest in it. It's not sellable...

What you do is replication and it's not real backup... With replication if something would be damaged on your primary data storage you'll have the same bad thing on the replica. So it's still make sense to run some backup software doing incremental backups.

[ed@zugzwang.com]

Hi Anton,
Thanks for the quick reply.
Now I understand.
Ed

[anton (staff)]

...and the worst thing: 1 GbE is capable of ~110-120 MB/sec STR and local HDD can do twice that much. Squadron goes always with the speed of the slowest ship so replica will come @ a very high price for you - writes being slowed down. Keep fast RAID0 or SSD (or RAID0 from SSDs like I do) as your primary storage and run a NAS/SAN box (RAID1 or RAID10 as I hate RAID5/6) for backup destination storage. That's a way to go. Backup application is of your choice. I run VDI so use VEEAM.

Hi Anton,
Thanks for the quick reply.
Now I understand.
Ed

[ed@zugzwang.com]

-Ahhhh......
Thanks for the additional explanation. i was assuming that the write would complete when the active device completed- and the passive device would just bufffer-up and run longer. Because in testing large file writes with local .img as active and iSCSI as passive.... the network traffic to the iSCSI kept going 15-20 seconds after the job ended. I'm fooled again
Thanks Anton,
Ed..

[anton (staff)]

That's b/c writes never (virtually...) touch the disk immediately. Writes are buffered in RAM by Cache Manager and flushed to disk by lazy writer @ background. That's why you see fast completion and so-called spikes of traffic on the network.

-Ahhhh......
Thanks for the additional explanation. i was assuming that the write would complete when the active device completed- and the passive device would just bufffer-up and run longer. Because in testing large file writes with local .img as active and iSCSI as passive.... the network traffic to the iSCSI kept going 15-20 seconds after the job ended. I'm fooled again
Thanks Anton,
Ed..