Hi,
This is relatively urgent.
We are facing the following problem.
We need to move data on FAS270C (appr. 2TB) onto another set of disks. There is a single aggregate that contains root volume and one or two (on two heads respectively) large volumes with single LUN on each. All volumes except root are 100% full. DOT 7.0.3.
Disks are 146GB 10K, RAID_DP, 14 disks on each head (average). Target set is 300GB 10K, RAID_DP, 11 disks on each head. I tested (under normal conditions) vol copy from root volume onto root volume copy. It tool 3 and half minutes for 440MB! It is absolutely unacceptable, we cannot afford this downtime.
No the questions are
- How are chances that without client load vol copy will be faster? What can explain that miserable performance for relatively small data set?
- Which speed can be expected under these conditions? Normal on-tape (non NDMP) backup from client systems - Windows only, SAN attach - runs approximately 20MB/s which are still acceptable.
- will aggr copy be faster and how much? If yes - how can I then switch root aggregate? To clarify - first I restrict target aggregate and make aggr copy. Then I put target aggregate online and get duplicate volumes, do not I? I _hope_ that filer will rename them, anyone can confirm?
- any chance updating to 7.0.5 *before* copy will help?
Thank you and I appreciate your answers. The downtime is scheduled for this weekend; it would quite a big problem with 2MB/s performance :(
Regards
Andrey Borzenkov Senior system engineer IT Product Services Fujitsu Siemens Computers Russian Federation
Telephone: +7 (495) 730-6220 ext. 247 Telefax: +7 (495) 229 91 30 Mobile: +7 (916) 678-7208 Email: mailto:Andrey.Borzenkov@fujitsu-siemens.com Internet: http://www.fujitsu-siemens.ru/
Have you checked the "option vol.copy.throttle" ? It should show be set to 10 for full volcopy speed.
-----Ursprüngliche Nachricht----- Von: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] Im Auftrag von Borzenkov, Andrey Gesendet: Freitag, 2. März 2007 10:53 An: toasters@mathworks.com Betreff: vol copy speed and copying over root aggregate Wichtigkeit: Hoch
Hi,
This is relatively urgent.
We are facing the following problem.
We need to move data on FAS270C (appr. 2TB) onto another set of disks. There is a single aggregate that contains root volume and one or two (on two heads respectively) large volumes with single LUN on each. All volumes except root are 100% full. DOT 7.0.3.
Disks are 146GB 10K, RAID_DP, 14 disks on each head (average). Target set is 300GB 10K, RAID_DP, 11 disks on each head. I tested (under normal conditions) vol copy from root volume onto root volume copy. It tool 3 and half minutes for 440MB! It is absolutely unacceptable, we cannot afford this downtime.
No the questions are
- How are chances that without client load vol copy will be faster? What can explain that miserable performance for relatively small data set?
- Which speed can be expected under these conditions? Normal on-tape (non NDMP) backup from client systems - Windows only, SAN attach - runs approximately 20MB/s which are still acceptable.
- will aggr copy be faster and how much? If yes - how can I then switch root aggregate? To clarify - first I restrict target aggregate and make aggr copy. Then I put target aggregate online and get duplicate volumes, do not I? I _hope_ that filer will rename them, anyone can confirm?
- any chance updating to 7.0.5 *before* copy will help?
Thank you and I appreciate your answers. The downtime is scheduled for this weekend; it would quite a big problem with 2MB/s performance :(
Regards
Andrey Borzenkov Senior system engineer IT Product Services Fujitsu Siemens Computers Russian Federation
Telephone: +7 (495) 730-6220 ext. 247 Telefax: +7 (495) 229 91 30 Mobile: +7 (916) 678-7208 Email: mailto:Andrey.Borzenkov@fujitsu-siemens.com Internet: http://www.fujitsu-siemens.ru/
Yes, it is set to 10. "man aggr" also mentions aggr.copy.throttle but this one does not exit.
-----Original Message----- From: Kießl Walter [mailto:kiessl@heidenhain.de] Sent: Friday, March 02, 2007 1:28 PM To: Borzenkov, Andrey; toasters@mathworks.com Subject: AW: vol copy speed and copying over root aggregate
Have you checked the "option vol.copy.throttle" ? It should show be set to 10 for full volcopy speed.
-----Ursprüngliche Nachricht----- Von: owner-toasters@mathworks.com [mailto:owner-toasters@mathworks.com] Im Auftrag von Borzenkov, Andrey Gesendet: Freitag, 2. März 2007 10:53 An: toasters@mathworks.com Betreff: vol copy speed and copying over root aggregate Wichtigkeit: Hoch
Hi,
This is relatively urgent.
We are facing the following problem.
We need to move data on FAS270C (appr. 2TB) onto another set of disks. There is a single aggregate that contains root volume and one or two (on two heads respectively) large volumes with single LUN on each. All volumes except root are 100% full. DOT 7.0.3.
Disks are 146GB 10K, RAID_DP, 14 disks on each head (average). Target set is 300GB 10K, RAID_DP, 11 disks on each head. I tested (under normal conditions) vol copy from root volume onto root volume copy. It tool 3 and half minutes for 440MB! It is absolutely unacceptable, we cannot afford this downtime.
No the questions are
- How are chances that without client load vol copy will be faster? What can explain that miserable performance for relatively small data set?
- Which speed can be expected under these conditions? Normal on-tape (non NDMP) backup from client systems - Windows only, SAN attach - runs approximately 20MB/s which are still acceptable.
- will aggr copy be faster and how much? If yes - how can I then switch root aggregate? To clarify - first I restrict target aggregate and make aggr copy. Then I put target aggregate online and get duplicate volumes, do not I? I _hope_ that filer will rename them, anyone can confirm?
- any chance updating to 7.0.5 *before* copy will help?
Thank you and I appreciate your answers. The downtime is scheduled for this weekend; it would quite a big problem with 2MB/s performance :(
Regards
Andrey Borzenkov Senior system engineer IT Product Services Fujitsu Siemens Computers Russian Federation
Telephone: +7 (495) 730-6220 ext. 247 Telefax: +7 (495) 229 91 30 Mobile: +7 (916) 678-7208 Email: mailto:Andrey.Borzenkov@fujitsu-siemens.com Internet: http://www.fujitsu-siemens.ru/
-
Borzenkov, Andrey -
Kießl Walter