I run a set scripts running under cron which make tar backups to an external
USB HDD.
Wednesday (7 July) night backup 22GiB took about 15min
Thursday (8 July) night backup 22GiB took 5hours 44min :-o
Nothing changed between these runs, the machine wasn’t rebooted, the USB
device wasn’t unmounted and remounted.
This has been a basic speed change - everything prior to 7 July runs at a
similar speed and everything since then is dead slow.
I’ve googled around and come up with things like …
Make sure the device is mounted async - I believe it is (options:
rw,defaults) but it wasn’t remounted in between.
Make sure that ehc1_hcd is loaded before uhc1_hcd - the entry in
/etc/modprobe.conf looks like it is and anyway there wasn’t a reboot
between.
Can anyone throw any light on this please - I dread to think how long a full
backup is going to take so I could definitely do with getting back to the
sort of speeds I used to get.
> Nothing changed between these runs, the machine wasn’t rebooted, the USB
> device wasn’t unmounted and remounted.
did you add or update any new software between the fast and slow
situation? (did you run any YaST Online Updates, or YaST > Software
Management, or zypper…at all?
have you tested the health of the USB mounted harddrive? how old is it
(i know of many cases of drives dying prematurely inside USB
enclosures)?
> I run a set scripts running under cron which make tar backups to an
> external USB HDD.
>
> Wednesday (7 July) night backup 22GiB took about 15min
> Thursday (8 July) night backup 22GiB took 5hours 44min :-o
Check some days more. Watch it happen.
Hard disk errors? Check the kernel or system log at that hour.
Busy system at the time?
–
Cheers / Saludos,
Carlos E. R.
(from 11.2 x86_64 “Emerald” GM (Minas Tirith))
> Fudokai wrote:
>
>> Nothing changed between these runs, the machine wasn’t rebooted, the USB
>> device wasn’t unmounted and remounted.
>
> did you add or update any new software between the fast and slow
> situation? (did you run any YaST Online Updates, or YaST > Software
> Management, or zypper…at all?
>
> have you tested the health of the USB mounted harddrive? how old is it
> (i know of many cases of drives dying prematurely inside USB
> enclosures)?
>
Since the sudden slow down I’ve rebooted and tried two different external
HDDs and a different USB port without any improvement.
Running hdparm -tT on the drive produces some silly figures for speed (looks
like an anaemic slug).
I didn’t run zypper or do any updates by hand but I don’t know when the
automatic updates kick in - YaST’s configuration module doesn’t say when it
runs - just ‘weekly’.
Fortunately the external HDDs I’m using are 1Tb HDDs in an Akasa caddy which
also has an eSATA connection so I’m now using that and speeds are back up
again to sensible values - what was that about greased weasel sh*t!
The server I’m backing up is on openSUSE 11.1 so I’ll now force myself to
upgrade to 11.2 (especially now I’ve got a full data backup) and then maybe
I’ll try USB again out of curiosity.
Interestingly enough, plugging the ext HDDs into my workstation produces
reasonable results from hdparm -tT - it’s only on the server it’s slow.
For the moment I don’t have the time to investigate further now I’ve got a
backup working Ok - I’ve got other things to keep me busy (in between
watching Le Tour).
> Since the sudden slow down I’ve rebooted and tried two different
> external HDDs and a different USB port without any improvement.
> Running hdparm -tT on the drive produces some silly figures for speed
> (looks like an anaemic slug).
…
> Fortunately the external HDDs I’m using are 1Tb HDDs in an Akasa
> caddy which also has an eSATA connection so I’m now using that and
> speeds are back up again to sensible values - what was that about
> greased weasel sh*t!
…
> Interestingly enough, plugging the ext HDDs into my workstation
> produces reasonable results from hdparm -tT - it’s only on the server
> it’s slow.
My Xtall ball says you have some other thing using the usb bus. The
bandwidth is distributed between all the devices there. Even a moving
rodent could slow down things.
Your disk is fine, as it works up to speed on usb on another computer,
or using eSATA on the server. Thus, it is the usb bus that is
problematic, as proved by trying another disk.
–
Cheers / Saludos,
Carlos E. R.
(from 11.2 x86_64 “Emerald” GM (Minas Tirith))
> On 2010-07-14 12:20 GMT Fudokai wrote:
>
>> Since the sudden slow down I’ve rebooted and tried two different
>> external HDDs and a different USB port without any improvement.
>> Running hdparm -tT on the drive produces some silly figures for speed
>> (looks like an anaemic slug).
>
> …
>
>> Fortunately the external HDDs I’m using are 1Tb HDDs in an Akasa
>> caddy which also has an eSATA connection so I’m now using that and
>> speeds are back up again to sensible values - what was that about
>> greased weasel sh*t!
>
> …
>
>> Interestingly enough, plugging the ext HDDs into my workstation
>> produces reasonable results from hdparm -tT - it’s only on the server
>> it’s slow.
>
> My Xtall ball says you have some other thing using the usb bus. The
> bandwidth is distributed between all the devices there. Even a moving
> rodent could slow down things.
>
> Your disk is fine, as it works up to speed on usb on another computer,
> or using eSATA on the server. Thus, it is the usb bus that is
> problematic, as proved by trying another disk.
>
Maybe so, but the only other thing connected via USB is a printer (HP
Laserjet 2200) which has been connected to the same port throughout :-o
> Carlos E. R. wrote:
>
> > On 2010-07-14 12:20 GMT Fudokai wrote:
> > My Xtall ball says you have some other thing using the usb bus. The
> > bandwidth is distributed between all the devices there. Even a
> > moving rodent could slow down things.
> >
> > Your disk is fine, as it works up to speed on usb on another
> > computer, or using eSATA on the server. Thus, it is the usb bus
> > that is problematic, as proved by trying another disk.
> >
>
> Maybe so, but the only other thing connected via USB is a printer (HP
> Laserjet 2200) which has been connected to the same port
> throughout :-o
Well… you can try to unplug it and retest the HD, and see what
happens…
What about the cable?
–
Cheers / Saludos,
Carlos E. R.
(from 11.2 x86_64 “Emerald” GM (Elessar))
> On 2010-07-16 13:04 GMT Fudokai wrote:
>
>> Carlos E. R. wrote:
>>
>> > On 2010-07-14 12:20 GMT Fudokai wrote:
>
>
>> > My Xtall ball says you have some other thing using the usb bus. The
>> > bandwidth is distributed between all the devices there. Even a
>> > moving rodent could slow down things.
>> >
>> > Your disk is fine, as it works up to speed on usb on another
>> > computer, or using eSATA on the server. Thus, it is the usb bus
>> > that is problematic, as proved by trying another disk.
>> >
>>
>> Maybe so, but the only other thing connected via USB is a printer (HP
>> Laserjet 2200) which has been connected to the same port
>> throughout :-o
>
> Well… you can try to unplug it and retest the HD, and see what
> happens…
>
> What about the cable?
>
Same cable to the printer - tried different cables for the HDD.
This is all moot now as I’ve switched to the eSATA connection and upgraded
the server to openSUSE 11.2 since the slow down.
It works Ok as above and I’ve not the time to delve deeper at the moment.
>DenverD wrote:
>
>> Fudokai wrote:
>>
>>> Nothing changed between these runs, the machine wasn’t rebooted, the USB
>>> device wasn’t unmounted and remounted.
>>
>> did you add or update any new software between the fast and slow
>> situation? (did you run any YaST Online Updates, or YaST > Software
>> Management, or zypper…at all?
>>
>> have you tested the health of the USB mounted harddrive? how old is it
>> (i know of many cases of drives dying prematurely inside USB
>> enclosures)?
>>
>
>Since the sudden slow down I’ve rebooted and tried two different external
>HDDs and a different USB port without any improvement.
>Running hdparm -tT on the drive produces some silly figures for speed (looks
>like an anaemic slug).
>
>I didn’t run zypper or do any updates by hand but I don’t know when the
>automatic updates kick in - YaST’s configuration module doesn’t say when it
>runs - just ‘weekly’.
>
>Fortunately the external HDDs I’m using are 1Tb HDDs in an Akasa caddy which
>also has an eSATA connection so I’m now using that and speeds are back up
>again to sensible values - what was that about greased weasel sh*t!
>
>The server I’m backing up is on openSUSE 11.1 so I’ll now force myself to
>upgrade to 11.2 (especially now I’ve got a full data backup) and then maybe
>I’ll try USB again out of curiosity.
>
>Interestingly enough, plugging the ext HDDs into my workstation produces
>reasonable results from hdparm -tT - it’s only on the server it’s slow.
>
>For the moment I don’t have the time to investigate further now I’ve got a
>backup working Ok - I’ve got other things to keep me busy (in between
>watching Le Tour).
>
>Alan
The other thing i have seen is the converter circuit card going south
on external drives. The drive and contents are fine, just the atapter
in the drive enclosure fails.
> This is all moot now as I’ve switched to the eSATA connection and
> upgraded the server to openSUSE 11.2 since the slow down.
>
> It works Ok as above and I’ve not the time to delve deeper at the
> moment.
>
> Thanks for your input anyway
Ok!
Now that the disk is on an eSATA port, don’t forget to run SMART tests
on it.
–
Cheers / Saludos,
Carlos E. R.
(from 11.2 x86_64 “Emerald” GM (Minas Tirith))
