nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Performance with Multiple Backingstores

Joost

Hi All,

Did any of you experience changes in the performance of your system after installing multiple backingstores?

We did some testing on our current system (one backingstore) to see if users on other continents could work in TeamSite without having to wait hours between clicks. Would it be wise to do this again when we have the new stores in place?

Assuming that the number of users and workload will stay the same.

Tanx

Joost

Find more posts tagged with

Comments

Migrateduser

We went to multi-store very shortly after an upgrade from TS5.0.x to TS5.5.2. The improvement in performance from the upgrade was so great it would have masked a reasonable loss when we added the stores. But we certainly didn't notice a performance hit from adding stores.

My experience only spans about 1/2 way across N. Am: Omaha,NE(svr)-Portland,OR and SF,CA(svr)-Austin,TX. But based on that, I strongly suspect you'll see more variance due to transcontinental network congestion than would be caused by server side impacts from multiple stores.

Still it would be wise, if it isn't too expensive, to do the testing again to attempt to quantify any such impacts. Simply by running a before and after test over the LAN you can isolate any multi-store impact from the effects of WAN latency.

Migrateduser

Here at Interwoven the engineering team switched from a single monolithic store to multiple stores when we converted from 5.0.2 to 5.5.1. We found a tremendous performance improvement at that time.

I agree looking for networking, or possibly disk caching issues may be good starting points.

lissa

james1

> Here at Interwoven the engineering team switched from a
> single monolithic store to multiple stores when we converted
> from 5.0.2 to 5.5.1. We found a tremendous performance
> improvement at that time.

But that probably had more to do with upgrading from 5.0 to 5.5 than with using more than one store...

-- James

--
James H Koh
Interwoven Engineering

nipper

>> Here at Interwoven the engineering team switched from a
>> single monolithic store to multiple stores when we converted
>> from 5.0.2 to 5.5.1. We found a tremendous performance
>> improvement at that time.
>
>But that probably had more to do with upgrading from 5.0 to
>5.5 than with using more than one store.

James, I am certain the all the sales critters out there are crying.

Are there any benchmarks of 1 store vs multiple ? I will happily ask management to spend the money if there is a performance increase.

Andy

Joost

Jep, me too!

13adam13

Joost,
You didn't say this is what you were trying to do, but I thought I would mention it to be sure.

Setting up TS to run in the Netherlands, and then a backing store in Asia, will not resolve network issues if the Asian users are hitting a TS in the netherlands.

TS locally mounts the stores, and all traffic is still served from the TS server.

This means that when a user in Asia makes a request to your TS server, the TS server would make a request to the store in Asia, the Asia store would respond to the TS server, and then the TS server would respond to the user.

Therefore, a local store will inevitably be faster than one on a WAN.

--Adam Wilson
Senior Programmer

Migrateduser

There is some documentation on the support site on MultiStore best practices, I'm not sure if you have seen it, check out:
https://support.interwoven.com/library/manuals/teamsite/552upgrade/us_006_multistore_bp.pdf

Regards,

lissa

Joost

Hi Adam,

We are going to keep the stores here, at least right next to the TeamSite server.

If we decide to setup a store in Asia and the US I think we will also put a TeamSite server over there. That would make more sense. But being Dutch I first want to test the cheaper way ;-)

Joost

BTW: Tanx Lissa! This is very helpfull