Disaster Recovery; APS deleting reports in Info Desktop 1

[Wolffen]

Hi All,
I'm in a bit of a bind. I was recently handed ownership of our Seagate Info server after the previous "admin" left. Monday morning, the server had some sort of crash. After rebooting, and launching Info Desktop, I found that most of the reports the company uses that were created in the last 8 months are missing in Info Desktop. Those reports that remain have lost all of their scheduling. I have been through a restore of the cinfo.mdb, all seagate files, and even a full disaster recovery to a new hard drive for the server (even with data as old as a month, or as new as the night before the crash). The results are the same. When I look in Info Desktop, virtually all instances, all the scheduling, and many reports are gone.
Here's where it gets more interesting: all the files that the cinfo.mdb should be pointing to are still on the hard drive. I can add them back to the Info Desktop by hand, and recreate the schedules if necessary. The files also show up in the ci_infoobjects table in cinfo.mdb as well. But they don't show in Info Desktop. Compact and repair on cinfo.mdb has had no effect. At this point, I'm at a loss. Has anyone else seen anything of this nature before? Can anyone offer any advice?
Our backups are all valid, but even with a full OS level restore, we get the same behaviour. We're running SI 7.0 on NT 4 (SP6a), updated virus scanner, etc. I'm at a loss at this point. I'm familiar with creating and editing reports, but I've never hit much in the admin arena of the system, and, since we're on an older version, Crytal will not offer any assistance. Any assistance board members could offer would be appreciated.

Thanks,

Landon Hobbs
DBA

 

[Wolffen]

Update: It's worse than I thought. While most of the reports are still under the si\input\reports folders, some are not, and others seemed to have reverted back to very old versions of themselves??? I've found several reports that I know I updated recently with time stamps from January, and all of my changes missing...Again, any assistance would be appreciated.

 

[shark963]

Not an answer to your queston - just to let you know it has happened to me on more than one occasion.
It was suggested by Seagate to compact and repair the database - but without any luck.

I did post a question similar to yours months ago - but I did not get any response.
I will be interested in any responses that you may get.

Good luck
B

 

[Wolffen]

Update: Turns out the missing reports were there, but another user had created similar reports ages ago that the system still pulled in. All the missing objects are in the CI_INFOOBJECTS table, but they aren't showing up in Info Desktop. I am able to add them back to the Info Desktop interface, but this basically creates a new file on the hard drive and points to that instead...Scheduling is pretty much gone, though. Very few reports had their schedules (5 out of 400), but some had the instance history of previous runs on the schedules that have helped me piece things back together.

One answer that could help me is this: what table in the CINFO database does Crystal store the schedule in? Is it the CI_RUNTIMEIMAGE table, or CI_SCHEDULE. I have a backup of the database from May 21st that has 3600+ records in the CI_RUNTIMEIMAGE table (whereas the current table has less than 500 records). If I can figure out why the objects disappeared from Info Desktop even though they exist in the CI_INFOOBJECTS table, I figure I can restore the schedules as well.

 

[Wolffen]

Update: Well, I wound up just restoring the May 21st CINFO.mdb file after compacting it, and everything worked like a champ. I lost any changes I made after 5/21, but that's better than losing all the scheduling and having to piece reports together. From what I've been able to piece together, the problem was that somewhere between 5/21 and 5/28 the CFINO.mdb file reached 1GB in size. Apparently, once the file hits 1GB, APS can no longer "talk" to it. All the reports I added, new schedules, etc. were kept in a buffer memory, I guess, and were wiped out once I rebooted. (Btw, it reached 1GB because, odds are, the db was never compacted since the system was put in place 3 years ago).

I'm in the clear now though. I just hate that I lost 3 days of
work to something this simple. Live and Learn, I guess.

 

[gsomerville]

Hello,

A stupid question, but are you sure you're connecting to the right APS?

If so, when a report is deleted from the Info desktop, the entries in the CINFO database are removed. (CI_INFOOBJECTS for the report and (CI_RUNTIMEIMAGE) any scheduled results or recurring schedules. It certainly sounds like a mass delete of the APS entries has occured.

If this information is missing, the only options I know about are to restore the database, or to manually add the reports back on the info server. ( right-clicking on a folder in the tree-view, selecting "New", "Report Object"
and browsing the source folder for the reports. The report should then be refreshed and the login details entered.)

 

[Wolffen]

We've only got one APS server here, so, yeah, I was connecting to the right one. For whatever reason, APS itself was doing the deleting, from what I could tell whenever I restored a backup of the 1GB (or compacted from 1GB) cinfo.mdb file. (I made the db read only before turning on the services at one point,and saw the event log fill up with errors complaining about not being able to delete records in ci_runtimeimage and ci_infoobjects). Once I go a compacted version of the db from before it hit 1GB, everything was kosher.

 

[OldSkoolSkater]

Hello

There is an advised size limit for cinfo.mdb, over which seagate recommend you migrate to an SQL db.
I can't remember what it is at the moment but it is pretty low - I will find out if required?

Surprising how much space a compact db frees up, eh?

Cheers

Steve