Unsolved
This post is more than 5 years old
9 Legend
•
20.4K Posts
0
1936
April 24th, 2016 12:00
TimeFinder/Clone is so broken in VMAX3
I am curious to hear from EMC why would you release TimeFinder/Clone functionality on VMAX3 when you know that is so broken ? When you create TimeFinder/Clone recreate on VMAX3, it performs some kind of scan and that scan can take hours depending on how big your environment is. For example i have a 30TB Oracle DB that consists of 86 x 500G volumes. Every "symclone recreate" takes 1.5 hours to complete even if there is absolutely no changes on source or target devices. My clone target is an R1 so until the clone becomes "Copied" i can't push it to my R2. You might say that i should be using TImeFinder/SnapVX, well it's just as broken Reproduce TimeFinder/Clone functionality on VMAX3 . Also the whole RDP issue with snapvx, unless you upgrade to a 3 weeks old Hudson, you can't see RDP utilization.
Feedback that i am getting from my local EMC team is that there is no desire/direction of fixing this TimeFinder/Clone issue, basically we are screwed. How can you let your decades long customers down like that ?
0 events found
LBM99
1 Rookie
•
119 Posts
1
June 30th, 2016 12:00
I should have updated this post a long time ago. I actually asked Dynamox if I could update and then forgot! Just want to let everyone know that this issue has a lot of attention and is not being ignored. Several of us met with Dynamox at EMC World and there have been follow-up emails and conference calls.
The TF design in VMAX3 places priority on other resources over the differential operations. However, this has likely gone too far in one direction. So while technically it is "working as designed" and no bugs were at play here, the behavior is something that needs to be addressed relatively soon. Engineering is looking at ways to speed up the actual copy process, and has fixes in test. The way the copy process is reported to SE / UVMAX is compounding the problem by making every differential operation appear to be full.
We are looking at short/mid/long term solutions, ranging from speeding up current code to new functionality. We have also been working with Dynamox to understand his process, environment, and business requirements to help us decide where our improvements should be focused.
Ayrton_Senna1
1 Rookie
•
62 Posts
0
September 13th, 2016 13:00
Hi dynamox . What's the current status of this issue ? Can you please explain the resolution if you received any from EMC.
Ayrton_Senna1
1 Rookie
•
62 Posts
0
September 13th, 2016 14:00
There seem to be a documented issue on this. Looks more of a reporting issue than operations to me.
472617 : VMAX3: TimeFinder Copy progress appears to be stuck. https://support.emc.com/kb/472617
dynamox
9 Legend
•
20.4K Posts
0
September 13th, 2016 18:00
Hudson (5977.813.x) + ePack (can't remember the number of head) is supposed to alleviate this issue. This "tracking" counter issues is supposed to speed up by 30% but it's still going to be there. Maybe Michael can give us an update on what else engineering is up to with this issue.