Hi,
Thanks again for the reply.
To answer your questions the snapshots were stored on my "C drive" which is an internally connectedl SSD drive (not USB) specifically as a Serial ATA (SATA) Marvell controller to a PCI Express 2.0 host, connecting a Intel 520 180GB SSDSC2CW180 drive. Not a faulty cable. This was coupled with a dediced physical disk consisting of dual drives on hardware raid for speed.
So no, I don't think its a case of a dodgy cable or a bad hardware setup.
Despite this reasonably good (at this point) hardware spec I somehow managed to tangle up the config so that VMWare was doing something incredibly inefficient because I/O was really slow performing VM (before I consolidated the snapshots) and as I say took about 6 hours + to merge/delete the snapshots that had been created.
I think the problem may have been that I had it set to create a snapshot whenever powering off and also had 'auto protect' enabled.
Now that I've 'deleted' old snapshots *(and turned off the auto-snapshot stuff above) the machine I/O is performing a lot faster.
Now what to do about tihs?
Firstly, these are not exactly obscure or hard to find options in the config and creating a snapshot sounds like a good thing that won't affect performacne except perhaps at the time it creates them. This is clearly not the case.
So yes I think it would indeed be good to warn me before doing something that may so drastically / negatively affect performance./
Specifically I think that it would be good if:
- there was some warning when you started to collect too many snapshots.
- the wording around 'deleting' snapshots was changed to be 'consolidate' or remove or something.. delete carries with it an implication of discarding rather than consolidating which is what it is actually doing.
- that, once again, even though it is hard to estimate exactly how long it will take it wouldn't hurt to give an indication that a long running operation may take a long time (as in longer than 10 mins) before you do it just so that a person can (for instance) decide to schedule the operation to take place outside of normal working hours.
Despite your assurance that VMWare does try to listen to customers the fact that this was first flagged as major issue back in 2007 and we still havent seen even the wording change gives me no confidence that VMware will actually respond to these concerns and address either the warnings or (preferably) the performance problems that lead to this happening in the first place.
Thanks though.
T