Slow backup performance

[mwd]

We have a few volumes getting backed on a server running extremely slow
throughput during the backup. If we create a new volume on the server, then
copy the data from an existing volume, it is MUCH faster on this new volume.
If you restore the same data from tape to the new volume, it is still slow.
So copying the data works, but just restoring the data does not. We'd like to
find the issue without have to recreate all of our volumes, if possible.
We're thinking something with the ACLs maybe, but can't pinpoint the exact
problem.

Any thoughts?

 

[Mathieu CHATEAU]

hello,

what is you underlying tape technology ?
Did you try to defragment the previous volume ?
Which software ? ntbackup ?


--
Cordialement,
Mathieu CHATEAU
http://lordoftheping.blogspot.com


"mwd" <mwd@discussions.microsoft.com> wrote in message
news2B65C71-1440-4655-8875-0392D986258D@microsoft.com...
> We have a few volumes getting backed on a server running extremely slow
> throughput during the backup. If we create a new volume on the server,
> then
> copy the data from an existing volume, it is MUCH faster on this new
> volume.
> If you restore the same data from tape to the new volume, it is still
> slow.
> So copying the data works, but just restoring the data does not. We'd like
> to
> find the issue without have to recreate all of our volumes, if possible.
> We're thinking something with the ACLs maybe, but can't pinpoint the exact
> problem.
>
> Any thoughts?

 

[mwd]

We are using an DLT-S4 (800GB/1.6TB) drive. We tried defragging the volumes
(this happens with all volumes on this server) but no improvement. We have
used both ntbackup and Brightstor Arcserve.

"Mathieu CHATEAU" wrote:

> hello,
>
> what is you underlying tape technology ?
> Did you try to defragment the previous volume ?
> Which software ? ntbackup ?
>
>
> --
> Cordialement,
> Mathieu CHATEAU
> http://lordoftheping.blogspot.com
>
>
> "mwd" <mwd@discussions.microsoft.com> wrote in message
> news2B65C71-1440-4655-8875-0392D986258D@microsoft.com...
> > We have a few volumes getting backed on a server running extremely slow
> > throughput during the backup. If we create a new volume on the server,
> > then
> > copy the data from an existing volume, it is MUCH faster on this new
> > volume.
> > If you restore the same data from tape to the new volume, it is still
> > slow.
> > So copying the data works, but just restoring the data does not. We'd like
> > to
> > find the issue without have to recreate all of our volumes, if possible.
> > We're thinking something with the ACLs maybe, but can't pinpoint the exact
> > problem.
> >
> > Any thoughts?
>
>

 

[Mathieu CHATEAU]

do you have only very small files ?
Is your controler's firmware and driver up to date ?

Is the DLT drive alone on it's own scsi card ?

--
Cordialement,
Mathieu CHATEAU
http://lordoftheping.blogspot.com


"mwd" <mwd@discussions.microsoft.com> wrote in message
news:0A7A2CB5-41D8-45F8-AB08-5A35535CB2F9@microsoft.com...
> We are using an DLT-S4 (800GB/1.6TB) drive. We tried defragging the
> volumes
> (this happens with all volumes on this server) but no improvement. We have
> used both ntbackup and Brightstor Arcserve.
>
> "Mathieu CHATEAU" wrote:
>
>> hello,
>>
>> what is you underlying tape technology ?
>> Did you try to defragment the previous volume ?
>> Which software ? ntbackup ?
>>
>>
>> --
>> Cordialement,
>> Mathieu CHATEAU
>> http://lordoftheping.blogspot.com
>>
>>
>> "mwd" <mwd@discussions.microsoft.com> wrote in message
>> news2B65C71-1440-4655-8875-0392D986258D@microsoft.com...
>> > We have a few volumes getting backed on a server running extremely slow
>> > throughput during the backup. If we create a new volume on the server,
>> > then
>> > copy the data from an existing volume, it is MUCH faster on this new
>> > volume.
>> > If you restore the same data from tape to the new volume, it is still
>> > slow.
>> > So copying the data works, but just restoring the data does not. We'd
>> > like
>> > to
>> > find the issue without have to recreate all of our volumes, if
>> > possible.
>> > We're thinking something with the ACLs maybe, but can't pinpoint the
>> > exact
>> > problem.
>> >
>> > Any thoughts?
>>
>>

 

[mwd]

The files vary from Word docs a few KB all the way to large images and ppts
several hundred MB.

I have updates all firmware, drivers, bios on the server (spend several
hours with HP verifying this).

The dlt drive is on it's own scsi card (ultra320).

Again, it works fine once you create a new volume and copy the data from the
old volume to it. just not on the original volumes or if you recreate a
volume and restore the data from tape.

thx.

"Mathieu CHATEAU" wrote:

> do you have only very small files ?
> Is your controler's firmware and driver up to date ?
>
> Is the DLT drive alone on it's own scsi card ?
>
> --
> Cordialement,
> Mathieu CHATEAU
> http://lordoftheping.blogspot.com
>
>
> "mwd" <mwd@discussions.microsoft.com> wrote in message
> > We are using an DLT-S4 (800GB/1.6TB) drive. We tried defragging the
> > volumes
> > (this happens with all volumes on this server) but no improvement. We have
> > used both ntbackup and Brightstor Arcserve.
> >
> > "Mathieu CHATEAU" wrote:
> >
> >> hello,
> >>
> >> what is you underlying tape technology ?
> >> Did you try to defragment the previous volume ?
> >> Which software ? ntbackup ?
> >>
> >> --
> >> Cordialement,
> >> Mathieu CHATEAU
> >> http://lordoftheping.blogspot.com
> >>
> >> "mwd" <mwd@discussions.microsoft.com> wrote in message
> >> > We have a few volumes getting backed on a server running extremely slow
> >> > throughput during the backup. If we create a new volume on the server,
> >> > then
> >> > copy the data from an existing volume, it is MUCH faster on this new
> >> > volume.
> >> > If you restore the same data from tape to the new volume, it is still
> >> > slow.
> >> > So copying the data works, but just restoring the data does not. We'd
> >> > like
> >> > to
> >> > find the issue without have to recreate all of our volumes, if
> >> > possible.
> >> > We're thinking something with the ACLs maybe, but can't pinpoint the
> >> > exact
> >> > problem.