Jump to content
Sign in to follow this  
arndwetzel

MacPro Xeon - 10.4.7 Server - Retrospect 6.1 Server crashes every time

Recommended Posts

Retrospect 6.1 does not work correctly on 10.4.7 Server Universal on Intel hardware due to a problem in the operating system. I don't think any experimentation is going to fix this.

Share this post


Link to post
Share on other sites

We have recently purchased an MacPro to use as a server, and are currently waiting for OSX Server. This thread has made me a little bit nervous...

 

As a temporary workaround: is it possible to install Retrospect Client on OSX Server 10.4.7 Universal, and backup harddrives from another Mac? Im not looking forward to manually backing up our archive every evening.

Share this post


Link to post
Share on other sites

All I can say is that with a Maxtor SATA drive rather than the Seagates that come in the MacPro, Retrospect is running fine for my customer. Not an easy fix, but a workaround that got me out of a hole...

Share this post


Link to post
Share on other sites

Backing up from network clients to a file on the internal SATA drives on the Mac Mini seems to work without any problems.

 

Attempting to backup FROM the internal SATA drive on the Mac Mini causes Retrospect to terminate.

 

I'm thinking of using an external Firewire drive instead of the internal drive as a workaround. Can anyone at EMC Insignia confirm whether the bug in question affects backing up from the Boot drive (regardless of type)? all internal SATA drives? or drives from particular vendors regardless of interface? Does the bug affect only 10.4.7 Server or any install of 10.4.7 universal on Intel?

 

It's a little frustrating when the officlal word is just "Retrospect 6.1 does not work correctly on 10.4.7 Server Universal on Intel hardware".

 

Thanks!

Share this post


Link to post
Share on other sites

Quote:


Does the bug affect only 10.4.7 Server or any install of 10.4.7 universal on Intel?

 


 

 

The bug only effects 10.4.7 Server (universal installer) on Intel hardware.

Share this post


Link to post
Share on other sites

Quote:

It's a little frustrating when the officlal word is just "Retrospect 6.1 does not work correctly on 10.4.7 Server Universal on Intel hardware".

 


 

Perhaps. But an accurate description of the issue would be "Retrospect 6.1 does not work correctly on 10.4.7 Server Universal on Intel hardware due to a problem in the operating system."

 

Now, it's possible that some operations might work, while others might fail. But such a condition would still be well described by the above statement.

 

Software developers generally try to find conditions that break the program, so they can fix it. Then it goes to SQA for various types of testing before any claim to stability is made. EMCInsignia is unlikely to suggest work-arounds unless and until such workarounds are tested. But their finite resources are probably going to finding a fix right now.

Share this post


Link to post
Share on other sites

May I suggest that anyone having this problem turn in a bug report to Apple via the RADAR system. The only way that Apple puts resources toward fixing a bug (and this is Apple's bug, not Retrospect's bug) is if it is seen to affect a lot of people. Apple might think that this is only an esoteric bug found by EMC/Insignia that isn't affecting anyone. But if it is affecting you, file a bug report so that its priority is raised for fixing.

 

If you are already registered as an Apple developer, go to

Apple RADAR login page

and log in and report the bug. If you are not a developer, sign up for a free developer's "online" account here:

Apple Developer Connection membership page

and then you will be able to log in and report the bug (and you will also have access to documentation, tools, etc.).

 

there's also an excellent article on AFP548.com on how to report bugs to Apple and how the system works:

The Art of Bug Reporting

 

That's the way the system works, and is how bugs get prioritized for fixing. And you will get notified when the bug is fixed, and you might even be asked to help test the fix.

 

Russ

Share this post


Link to post
Share on other sites

...or... EMC insignia can inform Apple about the problem (which they probably already have), and then we can just wait for the update.

 

Im not going to test if my server crashes because of Retrospect, neither am i going to report a bug that i (thankfully) havent experienced myself.

Share this post


Link to post
Share on other sites

hi emac,

 

Quote:

...or... EMC insignia can inform Apple about the problem

 


 

you mean like what Mayoff says here?

 

Quote:

A bug has been opened with Apple and they are investigating. We are doing our best to come up with a workaround, but due to the nature of the problem a workaround will not be as good as an actual bug fix in the operating system.

 


 

Quote:

Im not going to test if my server crashes because of Retrospect, neither am i going to report a bug that i (thankfully) havent experienced myself.

 


 

wow. thanks for your contribution. you are a real asset to the Macintosh community. what would we do without people like you? dohgrem.gif

Share this post


Link to post
Share on other sites

We have confirmation that Apple is aware of this bug and they are investigating it. I think they might have even been aware before we reported it to them.

Share this post


Link to post
Share on other sites

Thank you so much for your contribution, waltr. YOU really must be an asset to the Macintosh community. tongue.gif

 

Besides: it would be a lie to report a bug to Apple, if I was reporting someone elses bug. I dont mind being honest, in fact i cherish it.

 

My company pays good money for Mac hardware and software. My statement had nothing to do with contributing to the community (as you correctly pointed out), as it had to do with pointing out that Apple has to contribute to its customers.

 

If the problem doesnt lie with Retrospect software, then they shouldnt have to fix it.

Share this post


Link to post
Share on other sites

On 09/19/06 at 05:49 AM, emacoolit posted:

Quote:

Im not going to test if my server crashes because of Retrospect, neither am i going to report a bug that i (thankfully) havent experienced myself.

 


On 09/20/06 at 01:30 AM, emacoolit posted:

Quote:

Besides: it would be a lie to report a bug to Apple, if I was reporting someone elses bug. I dont mind being honest, in fact i cherish it.

 


 

Has anyone asked you to report a bug that you havent (sic) experienced, or to report someone elses (sic) bug?

 

On 09/15/06 at 07:55 PM, rhwalker posted:

Quote:

May I suggest
that anyone having this problem
turn in a bug report to Apple via the RADAR system.

 


 

emacoolit, here's a web page just for you:

web page

 

emacoolit, I so appreciated your contribution to the discussion because it made everything clear for me. Have a great day!

 

Russ

Share this post


Link to post
Share on other sites

Further investigation has shown that this issue is not isolated to 10.4.7 Server (universal) on Intel.

 

If you have ACLs enabled on 10.4.7 Workstation (universal) on Intel hardware, you may also see this issue. ACLs are typically off by default on workstations, so the problem is usually not encountered.

 

If you disable ACL's on 10.4.7 Server, you may be able to work around this crash. We have not tested this since ACL's are on by default with the server operating system. I wonder if this would explain why replacing of the hard disk would avoid the crash?

Share this post


Link to post
Share on other sites

Quote:

If you disable ACL's on 10.4.7 Server, you may be able to work around this crash.

 


 

Robin,

 

Does your testing indicate whether this only occurs if (one or some of):

 

(1) the volume being backed up has ACLs enabled in WGM on shared volumes?

(Sharing > General > Enable ACLs on this Volume)

 

(2) ACLs are enabled on some share, even if they are not being backed up?

 

(3) ACLs are enabled on a volume on which a Retrospect subvolume is defined?

 

(4) ACLs are enabled on the volume/share but no ACLs are defined?

 

I don't believe that ACLs are needed on the boot LUN (if you've got a separate boot volume, with user data on a different volume), only POSIX.

 

Russ

Share this post


Link to post
Share on other sites

1) I don't know the answer to this in relation to Shared volumes. I do know it will fail using the Retrospect Client.

 

2) If we are not backing it up, I suspect it won't be a problem

 

3) Subvolumes will have problems if the parent volume has ACLs turned on.

 

4) I don't think ACLs need to be defined. We reproduced just by enabling the option in 10.4.7 Workstation.

 

Just to make things more clear, this is Apple bug #4731251

 

We are investigating a workaround to the bug. If we are able to provide a workaround, I will post info in this thread.

Share this post


Link to post
Share on other sites

Quote:

The only way that Apple puts resources toward fixing a bug (and this is Apple's bug, not Retrospect's bug) is if it is seen to affect a lot of people. Apple might think that this is only an esoteric bug found by EMC/Insignia that isn't affecting anyone. But if it is affecting you, file a bug report so that its priority is raised for fixing.

 

Russ

 


 

Hence my statement: since I have read about the bug in this thread, I will not willingly risk crashing anything on my companys server. It IS affecting me, I just havent experienced it firsthand, and dont want to either.

 

Im sorry for those who have had some trouble, but I am glad that they inform other people about it. Shouldnt they be glad that people actually have been warned?

 

Offtopic:

I have now been accused of not contributing to "the community", and also been referred to an anger management site - which of course is (meant to be?) provacative...

Share this post


Link to post
Share on other sites

I have now installed Retrospect on our new Mac Pro Server, after disableing ACL for sharepoints. Retrospect ran just fine last evening. Thank you for your advice!!!

Share this post


Link to post
Share on other sites

Phil,

 

If you have a Server Operating System, then ACLs are turned on by default. You have to go out of your way to turn them off.

 

If you have a workstation operating system then ACLs are off by default and you have to go out of your way to turn them on.

 

If you really did not have ACLs turned on, then you are having a totally different issue.

Share this post


Link to post
Share on other sites

This might not be related to this thread, but Retrospect has been acting incredibly strangely on my Mac Pro running 10.4.7 (not Server). When I first installed it and set it up (I copied the /Library/Preferences/Retrospect folder from my old G5 to the Pro, as detailed in the manual for moving to a new computer, and updated to the latest version of Retrospect), it seemed to work fine. Then after a restart it failed to automatically run the backups. I found a blurb in the knowledge base that said that was a known bug. Then when ever I'd launch Retrospect, it would sometimes crash LaunchRetro or LaunchRetroHlper right after entering my password (I've always had the "Always require password" turned on). Now it *always* crashed after entering my password. So I can't even do a manual backup! This is most annoying.

 

Any ideas?

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×