Jump to content

Recommended Posts

Even though our OS X Server has not crashed, we have started getting an Error 505 (client is reserved) on the OS 9 computer that backs the server up.

 

 

 

The only resolution the help index suggests is to reboot both machines...but I can't keep rebooting the server just because Retrospect is having a problem.

 

 

 

The server does show up as a client, and that it's responding, but apparently Retrospect can'ty log into it to access the files.

 

 

 

Any suggestions? Or is this a bug?

 

 

 

Thanks!

Share this post


Link to post
Share on other sites

In reply to:

Even though our OS X Server has not crashed, we have started getting an Error 505 (client is reserved) on the OS 9 computer that backs the server up.

 

 

 

The only resolution the help index suggests is to reboot both machines...but I can't keep rebooting the server just because Retrospect is having a problem.

 

 

 

The server does show up as a client, and that it's responding, but apparently Retrospect can'ty log into it to access the files.


 

 

 

- What does the Status field of the OS X Client window say?

 

 

 

- Does rebooting the OS X machine make a difference?

Share this post


Link to post
Share on other sites

While rebooting the server may help, (and it does in my case) some servers can't be rebooted everytime Retrospect loses it's connection. I get the 505 message when I log out of the client server. If I start up a screen saver, I'm fine, but as soon as I log out, I'm done for and have to reboot.

 

-Alan

 

 

Share this post


Link to post
Share on other sites

The Client window shows the OS X Server and indicates "Responding," but if I try to see the volumes, it indicates that it can't log in.

 

 

 

Restarting the Server does resolve thew problem, but, as Alan said, I can't keep restarting the server because of that.

 

 

 

Is this a Retrospect bug? Would I be better off running the Backup from an OS X workstation?

Share this post


Link to post
Share on other sites

I wrote:

 

What does the Status field of the OS X Client window say?

 

 

 

To which Indy replied:

 

The Client window shows the OS X Server and indicates "Responding," but if I try to see the volumes, it indicates that it can't log in.

 

 

 

I should have been more clear. What does the Status field of the OS X Client Application (the program that runs on the remote OS X computer) say?

 

 

 

Dave

Share this post


Link to post
Share on other sites
Guest

We are having the exact same problem here. The client's status when this happens says its on.

 

 

 

We also seem to get error 515, however we know there are no network or communication problems. Currently we have to mount the Fileserver volumes on to the desktop of the os9.1 machine in order to back them up ** thus leaving important things like file permissions etc etc out of the loop. We really need to get this working the right way.

 

 

Share this post


Link to post
Share on other sites

In reply to:

We are having the exact same problem here. The client's status when this happens says its on.

 


 

 

 

The Status field is in the middle of the window, below the client's name, which is below the horizontal seperator line, which is below the Dantz logo, which is to the Left of the radio buttons that say "On" and "Off."

 

 

 

What does the Status field of the OS X Client Application (the program that runs on the remote OS X computer) say?

 

 

 

Dave

Share this post


Link to post
Share on other sites
Guest

I cannot give you the exact status... however I can tell you that it seems to lock it. Meaning whatever the last action we took on the server (with respect to the osx client) stays in the status.. For instance we removed the client from a daily backup... 5 days later the osx client still says In use by "Administrator" ....

 

even though it is no longer in use and the retro server has been closed and restarted. The only way to clear this is to 1. reboot the server which is just not practically or to terminal in and kill the process id for pitond.

 

 

 

On a side note:

 

 

 

This is what we are running:

 

Mac OS9.1 machine as the retro server

 

Max OSX (G4) as a file server <--errors everytime from 216, 505, 515. (This is a software raid... could that be the problem??)

 

Two Mac OSX workstations <---back up properly

 

Many Mac OS9.x machines, Win2k, and Win98 machines. <---backup properly

Share this post


Link to post
Share on other sites

I cannot give you the exact status...

 

That's a shame; it's pretty hard to think up a diagnosis without knowing the symptoms!

 

 

 

For instance we removed the client from a daily backup... 5 days later the osx client still says In use by "Administrator" ....

 

even though it is no longer in use and the retro server has been closed and restarted.

 

OK, that's helpful

 

 

 

The only way to clear this is to 1. reboot the server which is just not practically or to terminal in and kill the process id for pitond.

 

Not only is it not practical to reboot an entire Unix machine just to kill one process, it's entirely unnecessary!

 

 

 

You can kill the pitond process from the OS X Client application window by holding the Apple key and clicking the "off" button.

 

 

 

You can use the terminal to find and kill the process ID, and you can restart the pitond process from a shell session too. And you can easily connect via ssh to the machine and do all this from another machine if necessary.

 

 

 

On OS 9 clients, if network communication was lost the Retrospect control panel sometimes showed as "in use" until the entire computer was restarted. At least with OS X you can get control back of your machine without such drastic measures.

 

 

 

Max OSX (G4) as a file server <--errors everytime from 216, 505, 515

 

I can't find any reference to an error 216 in the user's guide...

 

 

 

This is a software raid... could that be the problem??

 

I don't know; do you loose network connections when you backup only the boot volume (since Apple's software RAID doesn't allow you to boot from the RAID, I presume you're using a seperate boot drive?).

 

 

 

What about the physical network differences between the OS X clients that work fine, and the OS X client that looses contact? Can you swap network wires, or swap ports on the switch, or other basic network troubleshooting? Being certain that there is no network communication problem sounds like the surest way to miss finding what the problem might be!

 

 

 

Dave

 

 

 

 

 

 

Share this post


Link to post
Share on other sites
Guest

<<>>

 

 

 

Thats a NO. You cannot.. you get a nice swirling rainbow !! You have to kill it in the terminal window. (under our circumstances anyways.. we have tried to kill it holding command (apple key) it just wont DIE!

 

 

 

 

 

the OSX server is a server.. on a 3com gigabyte switch.. It server hundreds of gigs of data with no problems at ALL. This is not a network communication problem, unless its in regards to the retrospect client protocal .. which we are sure is the problem.

 

 

 

 

 

<<<>>>>

 

 

 

No we lose connections while trying to back up the "boot" volume which is in fact interesting you mention because it seems when one finishes backing up we have a problem.. lets say the "RAID" backs up ok and moves on to the "boot" volume.. it dies.. says the client is reserved.. and vice versa..

 

 

 

My idea for possibly fixing it was defining "subvolumes" or something to back up bits of data at a time. the raid is 136 gigs!!

 

 

 

 

Share this post


Link to post
Share on other sites

In reply to:

the OSX server is a server.. on a 3com gigabyte switch.. It server hundreds of gigs of data with no problems at ALL. This is not a network communication problem, unless its in regards to the retrospect client protocal .. which we are sure is the problem.


 

Ah, I see, you know what the problem is.

 

 

 

But the server doesn't live in a vacum.

 

 

 

According to the Retrospect User Guide, 515 errors (which the original poster to this thread did not report) can be generated from the network communication of the application machine as well as the client machine.

 

 

 

Are there _any_ differences between the physical network topology of the happy OS X clients and the unhappy OSXS client?

 

 

 

The only times I've seen the spinning cursor of doom in the OS X Client application was when I physically disconnected the ethernet cable in the middle of a backup from the back of the machine running Retrospect ( I posted my tests on this Forum). I got a 519 error in Retrospect (I've never seen a 505 error) and the client kept thinking that it was in use for 10 minutes until it realized that it was alone. But if I clicked the "off" button (without a modifier key) during those 10 minutes I got the hang every time. When I pressed Apple+off I never got a beachball.

Share this post


Link to post
Share on other sites

We just had the problem replicate itself a few moments ago. The OS 9 backup computer crashed in the middle of the verification.

 

 

 

We now get the Error 505 (client is reserved).

 

 

 

On the OS X Server, the client indiciates it is "In use by (computer)". It's as though it thinks the backup is still going on.

 

 

 

I just turned the client OFF and ON again, and the backup now seems to be working.

 

 

 

 

Share this post


Link to post
Share on other sites

In reply to:

We just had the problem replicate itself a few moments ago. The OS 9 backup computer crashed in the middle of the verification.

 

 

 

We now get the Error 505 (client is reserved).

 

 

 

On the OS X Server, the client indicates it is "In use by (computer)". It's as though it thinks the backup is still going on.


 

 

 

Had you left the OS X Client alone for a while, you probably would have seen it go from "In Use" to "Ready" after 10 minutes or so.

 

 

 

Remember, if the application machine had simply lost contact with the client momentarily it would gone into a "net retry" mode in an attempt to complete the backup. The client has to stay aware for this, so it stays "connected" for quite a while in anticipation of being reconnected to the application machine.

Share this post


Link to post
Share on other sites
Guest

Acutally with similar situations.. we have found .. many hours later that the client is still "reserved" and the processor on the OSX server is spiked by pitond. to oh about 90%

Share this post


Link to post
Share on other sites
Guest

<<

 

 

 

According to the Retrospect User Guide, 515 errors (which the original poster to this thread did not report) can be generated from the network communication of the application machine as well as the client machine.>>>

 

 

 

Ok. <--Not a novice here. There is no difference in the topology. The difference is one is running OSX server and the others are running OSX workstation (well or plain old OSX). we have read the guide, we know what the errors means... its NOT on our end... Maybe someone from Dantz should be paying attention here because from a couple other users with similar setups, we seem to be having the same problem suggesting a client bug, not a network problem.

Share this post


Link to post
Share on other sites

In reply to:

There is no difference in the topology. The difference is one is running OSX server and the others are running OSX workstation


 

Well, there are of course other differences that are obvious, such as different network hardware, different physical ports on both the switch and the patch panel, etc. My question was intended to elicit information that might not be so obvious. We all realize that the software differences between OSX and OSXS is pretty small, just some additional Apple programs and services running. And since most people do not see this problem it's important to try and find out what's the same about the few posters who _do_ see this. Heck, even a bug needs a way to be reproduced...

 

 

 

- Is the OSXS machine on the same subnet as the OSX machines?

 

- Is the OSXS machine on the same switch as the OSX machines?

 

- Does the OSXS machine have multiple network interfaces enabled in the Network preference pane?

 

 

 

It's pretty obvious that Dantz does read these boards. IrenaS posts here all the time (often to ask users for more information that they don't offer in their original post, and sometimes to give updates on the status of known or new problems). But people who want help with their problems via a community board have to be willing to help each other. Otherwise you can call Dantz tech support and deal with the company directly (I have no idea if companies who charge for tech support give refunds if the problem turns out be a bug, but you could always ask 'em).

 

 

 

Dave

Share this post


Link to post
Share on other sites

We are experiencing the same problem here with an OS X (10.1.3) workstation client. Actually, we're experiencing a whole constellation of problems discussed in other threads. These are:

 

 

 

1) Retrospect Client spontaneously turns itself off. This appears to be initiated by the pitond process quitting. There seems to be no obvious pattern to this, but pitond won't stay running for more than a couple of hours. If Retrospect attempts access, we either get error 541, client not running, or error -1028, client not visible on network.

 

2) If we can manage to keep the client running until the backup of the client begins, the backup consistently fails after a few hundred MB have been backed up. The log indicates error 515, piton protocol violation.

 

3) If we attempt to reaccess, we then get error 505, client reserved. The client status on the client machine shows "In use by [backup script name]," and the pitond process is occupying 89-99% of processor cycles. The client status from the backup computer displays "Busy." It's remained in this state for 12+ hours, and I suspect it would stay that way forever unless pitond is killed.

 

 

 

This is our only OS X client at the moment, and the reason we updated to 5.0 from 4.3. We have tried the known "fixes:" making sure that ethernet is listed first in the OS X network ports; turning off speed threshold in the backup script; reinstalling the client. We have not yet been able to back up this machine since switching to OS X. It backed up fine when booted in 9.2.2. (The Retrospect Client control panel is disabled in Classic, so presumably there's no possibility of a conflict there.)

 

 

 

The OS X client is in a different subnet from the backup computer, which is running Retrospect Server 5.0.205 under 9.2.2. All of our clients have either a fixed address or a DHCP reservation, except for the ones in the backup computer's subnet. All regular MacOS and Windows clients have backed up flawlessly since we switched to 5.0.

 

 

 

This is very frustrating, and is the worst problem we've experienced with Retrospect in 10 years of use.

 

 

 

 

Share this post


Link to post
Share on other sites
Guest

Yeah we are having almost all those problems, including deferring a backup, yet it doesn't defer. We are on the same subnet here though. Still having all those problems. Are you running OSX or OSX Server on the client machine? We are having problems with Retro server running on a 9.2.2 machine, and the OSX server is running the client. but we have seen all the behaviors that you speak of... Look in your logs.. is there some other error before you get the 505 client reserved? We seem to get that error when one of the volumes on the OSX server gives a -43 error.. then when it moves to the next volume we get the 505 error.

Share this post


Link to post
Share on other sites

In reply to:

Had you left the OS X Client alone for a while, you probably would have seen it go from "In Use" to "Ready" after 10 minutes or so.


 

 

 

I've seen lots of Error 505s from OS X and OS X Server clients and I have never seen one spontaneously clear itself like this. The most common cause in our case is if Retrospect asks for new media in the middle of backing up an OS X or OS X Server client. After new media is inserted and erased Retrospect logs "Trouble reading files, error 515 (Piton protocol violation)" for the client that it was backing up and moves on to the next client. Thereafter the OS X client reports that it is reserved (error 505) until restarted. This happens with every OS X client we back up (if new media is required) and never happens with an OS 9 client. The workaround for us is to juggle the order the clients back up in to make sure the media breaks fall on OS 9 clients. Hardly a perfect solution...

Share this post


Link to post
Share on other sites
Guest

Furthermore, anytime there is ANY error, including a -43, on the OSX client, the client is automatically "not released" and the next time there is a 505 error (client reserved) on a workstation the solution is easy, not so easy on our OSX server which cannot be rebooted evertime Retro hangs it.

 

 

 

We even tried breaking the whole server down into subvolumes.. same problem. If we get a -43 on a certain subvolume, none of the other subvolumes after the one with an error back up because of the 505 error. This is not a problem with our network, this is a problem with RETROSPECT!

Share this post


Link to post
Share on other sites

Dantz is aware of and actively investigating this issue. The error 505 seems to come up after problems with the backup, mostly after the 515 (piton protocol) error.

 

 

 

Thanks for your patience. We will post here as soon as any further information is available.

 

 

 

Best regards,

 

 

 

Irena Solomon

 

Dantz Tech Support

Share this post


Link to post
Share on other sites

In reply to:

anytime there is ANY error, including a -43, on the OSX client, the client is automatically "not released"


 

But this doesn't happen for me!

 

 

 

Here's the test I tried:

 

 

 

- On OS X Client machine, define subvolume containing lots of files.

 

- From Retrospect application machine, start an Immediate Backup to a File Set of this subvolume.

 

- Pause the backup between Copy and Compare.

 

- From the client's Finder, remove a number of random files from the subvolume.

 

- Resume the backup.

 

 

 

In the log, each file that was removed was listed with an error -43.

 

OSX Client machine popped up dialog box listing the number of errors.

 

OS X Client application shows "Ready" status and the History field is current.

 

 

 

Retrospect 5.0.205 Mac OS X 10.1.5

 

Retrospect Client 5.0.528

 

 

 

 

 

- Troll, does this happen on each and every client on your network?

 

- Is every client that it happens to connected via the same network topology?

 

- Before you can insist that it's not your network have you done any tests to make sure?

 

- Easy test:

 

On a machine that consistantly shows the problem in the above test, connect the client to the Retrospect application machine via a cross-over cable and repeat the test.

 

 

 

Dave

 

 

 

 

Share this post


Link to post
Share on other sites

In reply to:

not so easy on our OSX server which cannot be rebooted every time Retro hangs it


 

Retrospect does not hang your OS X Server.

 

 

 

According to your narrative, the pitond process hangs.

 

 

 

There is no reason to reboot an entire unix machine just to kill/restart a single process. Do it from the Terminal.

 

 

 

Dave

Share this post


Link to post
Share on other sites
Guest

Dave,

 

 

 

I know you are trying to help, but this is what I am telling you. The only way to clear this problem is to reboot. You can kill the pitond process and restart it and you still get the same 505 error.

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  

×