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; } } }

TS backup server strategy

howie

Hello All-

I am working on a strategy to implement a backup teamsite server. We already have our backing store on a NetApp server, and would like to place things like /etc/passwd and all the Interwoven config files on this share as well. Both of our servers would then use symbolic links to point to this (the same) config. We want to avoid duplicating configurations and users across these two boxes. We would only have one server up at a time, so we don't see why this would not work. Has anyone tried to do this before? Are there issues I am not seeing?

thanks,
Howie

Find more posts tagged with

Comments

Migrateduser

What software are you using to do the backup and how are you accomplishing the hot swap?

jed

We already have our backing store on a NetApp server, and would like to place things like /etc/passwd and all the Interwoven config files on this share as well.

A symlink for /etc/passwd! Yikes. Keep the OS files local. That way, if the netapp goes down, the box will not die. For OS files, you are best of doing a nightly rsync or using nis/ldap or another password sync method.

howie

Hi-

Our content store is backed up seperately...What we are trying to do is provide a seperate server that we can quickly bring up in the event of a hardware failure on our primary. We are finding it is a pain to keep this backup box updated with all the users, custom scripts and configurations. In the event that our main box went down, we would like to be able to bring up the backup box and have it read all the same configurations as our primary server. The only difference would be the IP address.

thanks,
Howie

Migrateduser

Jed was right in that reading the same /etc/passwd file is bad design for a number of reasons -- how complicated would it be to change your authentication scheme? Also, my question remains: How are you going to manage the hot swap?

What you're asking is not trivial and might require some substantial work and unless there's something I'm not seeing right now, you may need to give some hard thought to how your users are authenticating themselves.

howie

Hi Dave-

The failover would be manual. Since the backup box would be configured identically to our primary server, it would just assume the identity of the primary box when we brought it up. We will probably use a VIP in dns so that the primary and backup have seperate IP addresses.

As far as setting up a seperate auth scheme, I am all for it. However it is my understanding that Interwoven requires every user on the system to have a unix system account.

-Howie

Migrateduser

If you were using a directory-based authentication scheme like LDAP, then your users would need to exist in that database which should be shared across these machines anyway.

How will people be accessing this machine, name or IP? Needless to say, the licenses are different based on machine name or IP and MAC ID. So, if you name server "knew" about a hot swap, it could route the requests accordingly and I think if the license is defined by IP, you should be okay.

I haven't tried this so this is a best-guess effort for me at this point.

howie

It would appear that even if we had our users in ldap, they would still need to be users on the local system. I know we can figure out a way to copy the users between the boxes. I am really wondering about all the other teamsite configurations. One big concern is the entity DB. I have seen a couple of references to placing this on a shared disk, but nothing that seems definitive.

I don't see why linking to common configurations would be an issue, as only one of these boxes will be up at any time. For those of you out there with a standby TS server, how do you handle things when and if you need to bring it up?

thanks,
-Howie