Posts Tagged ‘cache


ok, time to buy more server…

We are over loaded! Need more POWER!

Being citrix engineer I have one less worry, that is scalability. or provide so-called POWER on the fly.
These days, if you havn’t noticed Citrix is all about providing server on the fly.
Using XenServer, XenApp and Provision server and pump in the new server on the fly as required.

Or even manually, just build the server and put them on the farm. bang you got more power. (as long as you have space/hardware to put them but thats completely different topic of its own)

Ok, topic isnt about Citrix, its about Database.

Shit happens everyday in BACKEND world and one of the damn query is all about

my F***ing app is slow!! do something! or else!

Yes… sad fact is that lot of times, your Citrix server is performing just fine, or UNDER performing for some case and finger gets pointed at backend database. Now it gets ugly and become finger point excersice

God send! I just stumped on software called memcached which scales your database cache(read only) to distributed servers sitting in the datacenter (or your comms room)

Well only problem I can see is that it requires initial design change to implement properly but damn it scales so well that I think CLUSTERED bloddy over priced M$ SQL or ORACLE or DB2 has gets kick at the back.

I didnt know thats possible! or M$ SQL offers distributed DB!

well Im not trying to convince MS evangelist (dont get me wrong, 99 out of 100 MS SQL runs just fine) or trying to tell others to stay away from MS SQL in intial design but please can someone THING other posibility before the damn implementation when DB needs to be scaled?

Here is another example, not recently but close enough, my employer has decided to invest big buck and implement ITIL compliant management system (I wont name the product…) System does incident management, change management, problem management, config management, etc,etc

Looks great on the book but when hundreads of users were hitting the DB, it corrapsed in the initial run due to over loaded backend.

Hey I work in “INCIDENT” world, I cant follow up the incident when the system is down even it is SEVERITY 1.
Vendor who designed the system, project team who implemented the system, even admins who runs it never thought it would be this bad. (well may be they did but I dare not to ask them as I know answer would be this thing is shit and we should be running away from it)

Problem was like DB has no ability to scale, and in the end, customer accounts had to devided in to multiple DB instances that forces people to login to multiple systems.

Well I bet it costs more as more DB licenses , more OS, more storage, more backup and more DAMN MANAGEMENT costs. And lost of productivity due to login on to multiple system to track incidents. Priceless.

What a waste of money.

and what I heard from management? it was a great implementation. Well I dont need to tell you I eventually quit the job.

August 2020

Greyeye Tweets