Mempack and DLloadLib errors

[Skyzefawlun]

Hello,

 

I'm currently running Retrospect Multi-server 6.5.350 on a Dell PowerEdge 2600 running win2K server sp4, backing up to a Dell PowerVault 110T 320SDLT1. We initially installed v6.0 and later upgraded via the downloaded update. We're currently using the version 5.4.110 drivers, and up until recently had no issues at all. Nothing has changed as far as I know on the system in question, or any of the target sources. What I'm experiencing is very frequent "Mempack: total heaps = 18" errors showing up in the operations log within minutes of the application launching. I get maybe half a dozen or more per hour whether or not a script is executing. Upon trying to close Retrospect, I recieve this error: "DLloadLib: ULoadLibrary C:\Program Files\Dantz\Restrospect\patch.rpx failed os error 127" . The program will attempt to quit, but never does. It remains responsive, can move, resize, etc. but will not shut down. I have rebooted the server, and launched retro immediately afterwards and was able to quit out at that point before any script had run. Checking the log after this quit, I found more Mempack errors and another error 127. It would appear that the last few times we ran the scheduled backup, it hung while attempting to create a snapshot of the same source computer's volume. Since the last good backup, this target computer has been rebooted and sitting at a login. This appears to not be an issue as we have gotten several partial backups of one of the 3 drives on this machine, but it still hangs while creating a snapshot. So there it is. I have no idea if all three of these problems are related, or what exactly is going on. Any advice, or guidance would be greatly appreciated.

 

Specs again:

 

Retro Multi-server 6.5.350 (web upgrade from 6.0 install)

Driver version 5.4.110

Dell PowerEdge 2600 (3g of RAM) w/ PowerVault 110T SLDT1 320gig

Windows 2k server with SP4 installed and updated

Also running SQL server 2000

[Skyzefawlun]

A few new developments...

 

Well, same thing happened last night on the attempted backup. We're now thinking that the mempack's we're seeing in the logs are only there because of the reporting levels we've set and aren't worried about them. So this morning after killing off retrospect again, I attempted to run an immediate backup of the drive on the server that it was hanging on. Within seconds of executing it however, the program became unresponsive and I had to kill it again. So I grumbled for a second or two and went to launch retro once more. Much to my dismay it told me to complete my upgrade I would have to provide a license for my copy of the program. Done. It reverted back to what I believe was a 2 year old configuration (maybe where we were at before going from 6.0 to 6.5?) I saw Mayoff's reply to another poster about the same problem and opted to just rebuild everything from this point as it seemed to be performing better. I could now close the app without it hanging, or generating the "os error 127". So at this point I'm thinking "Great! It's running better then ever, and I've got an excuse now to weed out those silly files that have been forcing me to scroll through page after page of logs each day" I figured for good measure it might be worth trying to back up that drive on that server one last time. This time it started and ran just fine, until it got to the snapshot. At this point, just like during the regular nightlies, it hanged. So what I'm wondering now is what could cause it to hang during the snapshot of this drive? That and what does the "removable storage manager" have to do with it? That's appears to be the current target when it hangs. Any ideas? Is the "os error 127" going to come back and bite us somewhere down the road?