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; } } }

Teamsite Load Balancing

singh

We have Teamsite servers running 6.7.1 SP1 on solaris. Both boxes are in two data centers. Data store of the Active ( Primary ) Teamsite server is synched to the Fail over Teamsite server using SRDF ( Kind of Bit wise rsynch ).

WE have tested SRDF thing and is working fine. Like after synching all the data stores are visible in failover teamsite.

Now for the load balancer some checks needs to be made on teamsite instace before declaring it as dead and then switch the users to other failover instance.

I am thinking of this as those checks:

1. Iwwebd not running ?
2. Servlet engine not runing ?
3. Writing a custom jsp for the servlet engine which will actaully chk the data stores and other health checks and give the output back to load balancers.

So I need input on this whole scenrio if any body see some faults or better ideas.

Thanks In Advance.

Find more posts tagged with

Comments

Antharix

Since you are planning to use two different stores, it is essential that both the stores are in sync before the switch occurs.

Comming to checking of the Active server for availability, you may adopt the "option 3" or scripts that run at the backends. But it is essential that you check for the following

a) Iwwebd
b) servlet engine
c) The server itself
d) network connectivity

If the server can be restored with a iwreset or by restarting servlet engine, then you might just do that. If the Server is down and cannot be restored in real quick time, then switch over should be considered.

singh

Thanks for the reply .. i will create something and will share.

madhug

Hi,

We also have the similar kind of setup. You can use tools like VCS (Veritas).

You can give the list of process to monitor, when the process goes down the VCS will automatically switch to the secondary server.

Why are you using the load balance here?

Are you using the both data center servers simultaneously.

Regards,
Madhu

singh

Sorry guys this project got delayed for some time and now back..

We are doing this for a failover site .. like if one Data center goes down 3 dns will switch users to other DC.

Quick question here :

We have already tested this setup and is working fine. One issue is that I am not able to replicate Jobs in Teamsite .. so I am curious where are actaul jobs stored that we see under workflow tab ?

Thanks
Singh

singh

Anybody : Where is the Job DB in teamsite .. I am pretty sure not in datastore so has to be some diff custom DB for storing Job related information similar to events DB

ISCBorisB

Anybody : Where is the Job DB in teamsite .. I am pretty sure not in datastore so has to be some diff custom DB for storing Job related information similar to events DB

TS keeps Workflow Objects in the Backing Store (/iw-store/workflow Directory), format is proprietary.
The iwwfconvert CLT can be used to copy active workflow jobs from source to target backing store.

singh

Cool . thanks for the reply .. let met try this

gzevin

We have Teamsite servers running 6.7.1 SP1 on solaris. Both boxes are in two data centers. Data store of the Active ( Primary ) Teamsite server is synched to the Fail over Teamsite server using SRDF ( Kind of Bit wise rsynch ).

WE have tested SRDF thing and is working fine. Like after synching all the data stores are visible in failover teamsite.

Now for the load balancer some checks needs to be made on teamsite instace before declaring it as dead and then switch the users to other failover instance.

I am thinking of this as those checks:

1. Iwwebd not running ?
2. Servlet engine not runing ?
3. Writing a custom jsp for the servlet engine which will actaully chk the data stores and other health checks and give the output back to load balancers.

So I need input on this whole scenrio if any body see some faults or better ideas.

Thanks In Advance.

in general, the task of failover is not so trivial as it appears to be. Have you spoken to Interwoven regarding your setup to check if this is a supported path?

I dealt with a similar setup back in 2005 and was in contact with some key sale executives from IWOV, who provided me with all available best practice documents at the time. And it turns to be that it is not *so safe* to bring up the system should it fail on another box. Because the very reason of the failure could be the store, so you are risking to bring the system into the vicious circle.

just my 2c worth....

singh

Thanks for the input .. it is certainly worth of an input...

Fail over in our is surely not for data store that means if Data store got corrupted then we have to go for restore. But unfortunately we did faced issues where nfs got stuck or box hardware went down , i mean every thing else then data store corruption Smiley Happy

...

We have already tested the setup and it is working fine till now except that we are not able to replicate "jobs" stored in one teamsite instance to the other once.

coming to the another suggestion given to check /iw-store/workflow backing store .. i cant find any backing store there in may case Smiley Sad

singh

Got it .. i did :

cat /etc/defaultiwstore and it was on a diff directory so now let me try to convert it

gzevin

Got it .. i did :

cat /etc/defaultiwstore and it was on a diff directory so now let me try to convert it

it's anecdotal, in one of the installs I saw the admin having different store settings in 2 places. ended up with workflow being written to a different location. looks like you are having a similar issue?

singh

hurree . .it is resoved ..

what we did :

cat /etc/defaultiwstore or iwgetstore command ... to get the actaul location of "workflow" ds. and the bhoom... Smiley Happy

so another hurdle gone from load balancing .. and it seems to be in good shape now .. of course great help from this fourm .. thanks guys ..