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 File Corruption

System

My company has experienced major Teamsite file corruptions recently (We just launched our project two months ago.). The version we are using is 5.52 sp1. Interwoven support has given possible causes of file corruptions, such as improper shutdown. Unfortunately, those causes do not really help us. We are still experiencing file corruption, even though everything is working fine. Does file corruption happen very often with Teamsite? Is there any practical way to prevent this from happening?

Find more posts tagged with

Comments

Migrateduser

What type of corruption are you finding? Are you sure that backup, antivirus, defrag, etc. is disabled for the Y: drive and the drive with the backing store on it?

I have only seen one case where the backing store became seriously corrupted. The problem was a bug in Interwoven's 4.5.x virtual file system code that would cause the NT server to bluescreen under certain conditions (Interwoven blamed Microsoft (OS), IBM and 3Com (hardware/drivers), sunspots, the devil and anybody/thing else they could think of before admitting it was their problem. Microsoft suggested that if we uninstall TeamSite the machine would stop bluescreening, which I have to agree with. It took a few months for Interwoven to deliver a patch, in which time we had to rewrite all the workflows, custom menu items, etc. to not use the feature that would cause the bluescreen). I have not seen any serious corruption (data loss) since the patch, but iwfsck will frequently report corruption, and Interwoven has a nasty habit of incorporating patched bugs from previous releases into subsequent releases (I claim this comes from maintaining the source code for TeamSite in TeamSite - I hope they don't really do that anymore). Running the tools to fix the corruption can cause data loss under certain conditions, such as current and/or old versions of files to be lost, but frequently it will do no damage (such as when you fix corruption on a directory). I do not know what causes Interwoven to report the corruption, or why the backing store doesn't just fix itself when it detects corruption.

I doubt you will be able to get much information on this issue, but please post whatever you find on this thread.

Migrateduser

I am still debugging this issue. I have one instance that I received errors the first time I ran iwfsck. I saw no errors after I ran it the second time. Interwoven support said I need to do "iwfreeze" in order to get accurate reports. I have heard from other people that you should run iwfsck multiple times to guarantee no errors exist.

It seems to me that the Teamsite database is really fragile. Shouldn't a database make sure all transactions are ¡§committed¡¨ properly and take care of system crash gracefully?

Migrateduser

It actually does do a decent job of this (for instance, if you reboot the server when huge submit/get latest/etc. operations are running, it will recover on boot), but I think there must be unrecoverable cases.

It is true that it will report errors that don't exist if the store is not frozen, and that most of the errors can be safely ignored (though you should fix them). I would not be too worried about the errors you are seeing, but I would ensure they are fixed, and check the files that they report on to ensure data is not lost.

You are pointing out what I consider to be a major architectural flaw in the product - a filesystem does not make a very good object store. So many things built around this are hacks - DCR file vs. output file, totally flat attribute storage, etc. I think it also causes performance issues. But if you had to pay $X to license a DB to run Interwoven then you would have less money to give Interwoven for TeamSite so they left that out. Database vendors do provide more reliable products, and they are the experts in that area (which is another reason I would never implement datadestroy - you will notice that a huge number of devnet posts are about weird DD errors).

I would say that Interwoven products in general are all very fragile.

ProfessorX

HELLLLLP
The Store is there but it is not being found.
Y: Is mounted but no dirs under it
Log in to teamsite but all I see is the the name and icon of the server in the WA

WIN2K TS 5.5.2

Here part of iwtrace

TeamSite I18N initialization successful!
[Thu Aug 14 13:07:25 2003]
*** IWServer Locale is --> English_UnitedStates.MS1252@Binary
[Thu Aug 14 13:07:25 2003] [Thu Aug 14 13:07:25 2003] Watch Event Thread started
TeamSite Server Starting.
[Thu Aug 14 13:07:25 2003] iwserver: 5.5.2 Build 9275 Interwoven 20020617
[Thu Aug 14 13:07:25 2003] Reading users and groups...
Local system: "WUCMS01"
[Thu Aug 14 13:07:25 2003] done. 23 users and 8 groups read.
[Thu Aug 14 13:07:26 2003] Cache size: 30000
[Thu Aug 14 13:07:38 2003] Store E:\iw-store\default: 1011 objects found in the log file.
[Thu Aug 14 13:08:35 2003] TMetafileIoEngine::GetHeader: Unable to read index 0x0:No such file or directory
[Thu Aug 14 13:08:35 2003] TModification::Commit: WritePoint failed: object ID = 40000, data size = 332
[Thu Aug 14 13:08:35 2003] Store E:\iw-store\default: recovery failed, store is disabled.
[Thu Aug 14 13:08:35 2003] TEnvironment::dtor found 0 dirty.
[Thu Aug 14 13:08:35 2003]: TNetworkShares::Add Can't NetShareAdd, parameter error 0x0
[Thu Aug 14 13:08:35 2003]

\Program Files\Interwoven\TeamSite\etc\iw.cfg: rpc_threadcount=64
[Thu Aug 14 13:15:57 2003]
*** IWServer Locale is --> English_UnitedStates.MS1252@Binary

The Professor-

Migrateduser

How are you running iwfsck?

ProfessorX

ran iwfsck

returned fine with no errors

The Professor-

Adam Stoller

What do you see when you run iwstat ?
What do you see when you run iwstoreadm -l ?

--fish
(Interwoven Senior Technical Consultant)

ProfessorX

Update--

I replaced iwstore form backup...it seemed to come up fine....I have been kind of lax on my CMS administration duties....I'll try and keep and eye on it.

Thanks for your help peoples

The Professor-