Jump to content

Retrospect Client hangs on Maxtor 5000 backup


randyray

Recommended Posts

Here is my story. Hope someone has seen this problem and has a fix for it.

 

I am using Retrospect 6.5 full version (not Express that came with Maxtor drive)

I have a backup machine (used for backups only) and it has 4-Maxtor 5000xts (USB) and a DLT drive. It runs Windows XP Pro, 1.8ghz, 512Meg mem. We'll call this machine "backup"

 

It is connected to a client machine on a local ethernet (100mbit). The client runs 6.5.127

 

version of client. This machine has 2-Maxtor 5000xts (USB) that hold a lot of large video files

 

(some as large as 22Gb). We'll call this machine "client"

 

I have the backup machine run an "incremental" every night to the client and backup the client Maxtors to the backup Maxtors. Most times all goes well. I've been running this configuration all summer. We'll call this prodedure "Maxtor backup"

 

Once a week I also run an incremental on the backup to the client to DLT tape. Most time this

runs well. We'll call this procedure "DLT backup".

 

While running the backups, I also have it compare the files it backs up.

Here is the failure I get. It happens most on the DLT backup, but I have had it happen on the

"Maxtor" backup.

 

Always during the compare phase and always on large files and especially when Retrospect calls for a next tape (compare phase) during a DLT backup, the backup machine will time out saying "Trouble reading files, error -519 (network communication failed)". If I go look at the client machine, the client is still running and the Maxtor drive it is comparing is still flashing its activity light (red). After a while (sometime 20 minutes), it stops. It happens on different

drives, etc.

 

Here is what I have observed. During a compare (once I have given the DLT drive the next tape), the Retrospect client software is running something on the Maxtor (and is tied up doing this).

 

The Retrospect host times out (network failure) waiting for it to respond. After the failure,

and while the Retrospect is still making the "activity light" flash on the Maxtor drive, I kill

the client process and the activity stops. If I don't kill the process, the client is tied up

until it "finishes" with the Maxtor drive (sometimes as long as 20 minutes later), but by this

time the host Retrospect on the backup machine has long since timed out.

I notice that this drive activity happens all the time but on smaller files, the client finishes

before the host has timed out and the compare continues.'

 

Questions:

 

- What is the client doing to the Maxtor drive that ties it up and keeps it from repsonding to

 

the host?

 

- Would I be better off accessing these large drives with large files directly from the backup

 

machine (as network drives) and bypass the client?

 

This is the ONLY problem I've had with this fine product and I would like to have some kind of

 

workable solution.

 

Thanks for any help.

If you need more info, please email me at

vendor@randyray.com

 

Randy

 

Link to comment
Share on other sites

Hi

 

Do you have Windows XP system restore enabled on the Maxtor drives? If so turn it off.

 

We should try to narrow this down to one device if we can...

Does the failure happen regardless of how many files have backed up? The client machine isn't going to sleep or some power saving mode is it? Is Autoprotect running during the backup?

 

Nate

Link to comment
Share on other sites

I didn't know about the problem with Restore and turned it off on all Maxtor drives...both the client ones to back up and the host ones that I back up to (backup set). Not sure about Autoprotect.....please clarify......I do already have Norton "protect" turned off for these drives. I'll check for a day or two and post results.

 

The failure always happens when I have large amount of data in the "compare" phase. Most of the time it is less than 10 files.....but those files can range anywhere from 5Gb to 22Gb each.

 

Thanks for the help....I'll post in a day or two.

Randy

Link to comment
Share on other sites

Thanks Nate.

I have turned off the "Restore" stuff for each of the Maxtor drives. I had already turned off the Norton "auto protect" feature for each of them.

The client machine is never allowed to go into sleep or any energy saving mode.

 

The files that seem more to failure are not large in qty...but large in size.

 

I have done 3 backups for very large files from the client machine to the host (to backup sets on the other Maxtor dives)....but this does not fail a often as the backup to DLT. I will try the DLT backup this weekend and post my results on Monday.

Randy

Link to comment
Share on other sites

Ok Nate here is the latest...and I'm out of ideas on what other info to gather. Maybe there is some debug switch I could turn on somewhere to gather more data.

It failed on a DLT backup.

Here is the most detail I can gather so far.

 

Same systems described in original note.

I am trying to backup a 25Gb file located on a Maxtor 5000XT (250Gb 5400rpm USB2.0) on the client machine across a local LAN to my backup (host) machine and the DLT drive locally attached to it.

The backup will cross 2 DLT tapes (last part of 1st tape and first part of second tape).

I start the backup and Retrospect starts the storing of the file onto the 1st DLT tape.

When the tape is full, Retrospect asks for the next tape in the backup set.

Before I put the tape in, I go and look for any activity on the client machine and Maxtor drive....there is none.....the Retrospect client window shows that the client is working on backup (not idle)...as I would expect.

 

Once I put the next tape in and it becomes "ready", the host machine waits on the client machine (for 20 to 30 minutes)...and does nothing else.

While it is waiting, I go and look closely at the client machine. The Maxtor's activity light is flashing with activity. When I look closely at what process (in Win Task Mgr) is causing the READ I/O activity, I see the following:

 

SERVICES.EXE has some READ I/O activity....I expect this.

LSASS.EXE has some READ I/O Activity, but not much.

PCPDS.EXE has a ton of I/O activity......I/O READs at 126/sec...or more

 

After a search, I find that PCPDS.EXE is a file in the Retrospect folder.

 

I continue to wait the nearly 30 minutes and sure enough, eventually the backup continues.

 

I've seen this before.

 

Now that the backup is finished, I had selected to compare the data.

The data compare starts.

It asks for the 1st tape and when ready, starts the compare.

 

(here's where the failure almost always happens)

 

When Retrospect asks for the next tape to continue the compare, I go and look at the client machine again....no activity...even with PCPDS.EXE.

When I give it the next tape and it comes "ready" it says "Comparing....."...but nothing happens.

I go to check the client machine and the Maxtor activity light is flashing activity, and PCPDS.EXE is performing a lot of READ I/O.

 

I wait at the host machine and notice it has changed from "Comparing...." to Retrying..."......for about 5 to 10 minutes....then fails with "Trouble reading files, error -519 (network communication failed)"

It immediately starts a small backup operation that has been waiting to go check on a few files on the same client machine and it fails with "Can't access volume My Photos on LABDELL, error -505 (backup client reserved)"....which I would expect since the Retrospect client is busy doing I/O like crazy to the Maxtor.

 

I go look at the client machine and it is still flashing the Maxtor light and PCPDS.EXE is still performing large amounts of READ I/O.

 

Since the compare part of the backup has already failed anyway and to try to prove that PCPDS.EXE is actually the app that is making the MAXTOR light flash, I kill the process and immediately the Maxtor light stops.

 

This is the failure scenario I almost always see.

 

My questions are:

 

1. Why is PCPDS.EXE doing all of this I/O before the next tape can continue? Is it reading down the 25Gb file to relocate the point it left off at or something like that?

2. I feel it is a "race" condition that when large files are backed up and they are compared across two tapes, the host "times out" (retry) waiting on the client....but why?

3. I haven't tried to see if bypassing the client (accessing the Maxtor on the client machine as a network drive on the host) would be a "work-around", but I may try this later.

4. Do you have any other ideas?

 

Thanks,

Randy frown.gif

Link to comment
Share on other sites

Hi

 

Hmmmmm, this is a new one for sure.

Can you run a test backup where you don't backup data from the maxtor drives? (i.e. move some large files over to the C drive and try the backup again)

 

Make sure you also get the latest driver update from the Dantz Website. Basic stuff - but have you uninstalled/rebooted/reinstalled the client software? Any chance you could do the same with the Network card on your client machine?

 

Nate

Link to comment
Share on other sites

  • 1 month later...

Nate,

Have not had time to try files from non-USB drive as your last post suggested.

I found a "work-around" in that I made the USB- Maxtor drives network drives to the backup machine. This bypasses the client on the "usb maxtor" machine. The delay is still there, but there is no failure. I will pursue when I have more time.

Randy

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...