Start a Conversation

This post is more than 5 years old

Solved!

Go to Solution

2089

November 21st, 2011 06:00

what is write pending slot in DMX?

how the slots are divided. If it is based on cache size, what is the standard slot size.

1.3K Posts

November 22nd, 2011 05:00

If the raw capacity is 512GB, my statement still applies.  There is cache used for meta data in the Symmetrix.  The amount of meta data used depends on the amount of storage capacity configured in the system.  What is left over for user data is your "available cache slots"

The maximum write pending slots is a percentage of the total user cache, usually 80%, but in 75 code is now 75%.

The device write pending limit is set to 5% of the writable cache. 

The DA write pending is obsolete, and even when it was used, was very rarely invoked.

The write pending limits are the point where any host writes must wait for a slot to be destaged before a new write can be allowed into the array.

2 Intern

 • 

1.3K Posts

November 21st, 2011 06:00

in 5773 code you can see 80% of total cache slots are available for system WP slots while 5774 or 5775 you could see it is 75% of total; look at `symcfg list -v` o/p

1.3K Posts

November 21st, 2011 06:00

DMX3/4 and VMAX always use a 64k cache slot for allocations.  So any write to cache will occupy a minimum of 64k. 

148 Posts

November 22nd, 2011 03:00

Cache Size (Mirrored)            :  262144 (MB)
# of Available Cache Slots       : 1578161

Max # of System Write Pending Slots  : 1263761

Max # of DA Write Pending Slots  :  631880

Max # of Device Write Pending Slots  :   63185

if the cache slot is of size 64K, then total available cache slots would be rounded upto 1578161*64K=~96GB.

So what about remaining cache(256GB is total)

148 Posts

November 22nd, 2011 05:00

Here the raw capacity is 512GB and i mentioned the mirrored cache of size 256GB. So my question is if 96GB is used as cache slots , what about remaining cache.

And any body can explain me what is the actual meaning of by system,DA,Device write pending limits and how they are inter related.

Cache Size (Mirrored)            :  262144 (MB)
# of Available Cache Slots       : 1578161

Max # of System Write Pending Slots  : 1263761

Max # of DA Write Pending Slots  :  631880

Max # of Device Write Pending Slots  :   63185

108 Posts

November 22nd, 2011 05:00

Enginuity uses Tag Based Caching (TBC) Least Recently Used (LRU)

algorithm for cache management. Enginuity divides global memory

into groups of several hundred slots. In the TBC data structure, four

bytes represent each slot. The four bytes contain information about

the last time the system most recently accessed this slot, whether the

slot is write pending, and other attributes. Each memory slot is

represented by one

tag. Tags are clustered into groups called extents

.

The TBC LRU algorithm determines which slot in an extent was least

recently used. This TBC LRU slot then loses its association with the

track/data that is stored. Getting a new slot means reading all the

presenting bytes of a TBC extent and choosing the oldest one to be

replaced. Changing one bit removes it from the LRU pool of available

slots.

All the CPUs in a Symmetrix system access all the TBC extents.

Enginuity manipulates the TBC extent under lock. To avoid

contention over one TBC extent, each CPU in the Symmetrix system

accesses the TBC extents in a different order, which guarantees even

loads on the different TBC extents at any given time. TBC simplifies

LRU implementation

1.3K Posts

November 22nd, 2011 05:00

Cache in DMX3/4 and VMAX is mirrored, so the total protected cache is 1/2 the raw cache.

The difference between 128GB protected cache and the 96GB is the cache used for meta data.

148 Posts

November 22nd, 2011 07:00

So is it kind of thing, that if the box is created with more number of devices then that will decrease the available cache slots.Because much of the cache is used for metadata info,which in turn reduce Box performance.

1.3K Posts

November 22nd, 2011 07:00

The meta data is essential for operation.  The decrease in cache may have an effect on performance, but maybe not.

No Events found!

Top