From: Richard S. <rs...@pa...> - 2003-01-22 18:44:03
|
Hi, While there is an alternative to NetBench under development (or at least being talked about as part of SNIA and SPEC), there is also a need to be able to run NetBench. What I would like to do is to automate the running of NetBench and script it so that data collection can be driven from our existing UNIX-based infrastructure. This would be for checking for performance regressions etc. Has anyone looked at whether or not NetBench can be scripted in any way? Please reply to me directly, as I am not on the list. Regards -- Richard Sharpe Software Engineer, Panasas Inc Pioneering Smart and Infinitely Scalable Storage Networks www.panasas.com, rs...@pa..., rs...@sa..., www.richardsharpe.com |
From: Cliff W. <cl...@os...> - 2003-01-22 19:33:52
|
> Hi, > > While there is an alternative to NetBench under development (or at least being talked about as part of SNIA and SPEC), there is also a need to be able to run NetBench. > > What I would like to do is to automate the running of NetBench and script it so that data collection can be driven from our existing UNIX-based infrastructure. This would be for checking for performance regressions etc. > > Has anyone looked at whether or not NetBench can be scripted in any way? > > Please reply to me directly, as I am not on the list. > > Regards > -- > Richard Sharpe > Software Engineer, Panasas Inc > Pioneering Smart and Infinitely Scalable Storage Networks > www.panasas.com, rs...@pa..., rs...@sa..., www.richardsharpe.com > > > ------------------------------------------------------- > This SF.net email is sponsored by: Scholarships for Techies! > Can't afford IT training? All 2003 ictp students receive scholarships. > Get hands-on training in Microsoft, Cisco, Sun, Linux/UNIX, and more. > www.ictp.com/training/sourceforge.asp > _______________________________________________ > Lse-tech mailing list > Lse...@li... > https://lists.sourceforge.net/lists/listinfo/lse-tech > We've just started looking at it, haven't thought about scripting. We're interested in doing something like that cliffw |
From: Richard S. <rs...@pa...> - 2003-01-22 19:45:14
|
On Wed, 22 Jan 2003 11:33:37 -0800 Cliff White <cl...@os...> wrote: > > Hi, > > > > While there is an alternative to NetBench under development (or at least being talked about as part of SNIA and SPEC), there is also a need to be able to run NetBench. > > > > What I would like to do is to automate the running of NetBench and script it so that data collection can be driven from our existing UNIX-based infrastructure. This would be for checking for performance regressions etc. > > > > Has anyone looked at whether or not NetBench can be scripted in any way? > > > > Please reply to me directly, as I am not on the list. [Deletia] > We've just started looking at it, haven't thought about scripting. > We're interested in doing something like that > cliffw OK, having spoken to a bunch of people, it seems that this is not possible. Another approach I am thinking of is to calibrate cifs_load_gen against NetBench, and then use cifs_load_gen to run regression tests, and re-calibrate from time to time. Regards -- Richard Sharpe Software Engineer, Panasas Inc Pioneering Smart and Infinitely Scalable Storage Networks www.panasas.com, rs...@pa..., www.richardsharpe.com |
From: Andrew T. <hab...@us...> - 2003-01-22 19:54:26
|
The only thing I was able to do was script the clients to start their=20 client.exe junk. I could not find any way to script the controller. Yes= ,=20 running NetBench can be a real PITA.=20 NetBench also requires MS Excel to process the results, another reason I = am=20 glad there is an alternative under development. =20 -Andrew On Wednesday 22 January 2003 13:33, Cliff White wrote: > > Hi, > > > > While there is an alternative to NetBench under development (or at le= ast > > being talked about as part of SNIA and SPEC), there is also a need to= be > > able to run NetBench. > > > > What I would like to do is to automate the running of NetBench and sc= ript > > it so that data collection can be driven from our existing UNIX-based > > infrastructure. This would be for checking for performance regression= s > > etc. > > > > Has anyone looked at whether or not NetBench can be scripted in any w= ay? > > > > Please reply to me directly, as I am not on the list. > > > > Regards > > -- > > Richard Sharpe > > Software Engineer, Panasas Inc > > Pioneering Smart and Infinitely Scalable Storage Networks > > www.panasas.com, rs...@pa..., rs...@sa..., > > www.richardsharpe.com > > > > > > ------------------------------------------------------- > > This SF.net email is sponsored by: Scholarships for Techies! > > Can't afford IT training? All 2003 ictp students receive scholarships= =2E > > Get hands-on training in Microsoft, Cisco, Sun, Linux/UNIX, and more. > > www.ictp.com/training/sourceforge.asp > > _______________________________________________ > > Lse-tech mailing list > > Lse...@li... > > https://lists.sourceforge.net/lists/listinfo/lse-tech > > We've just started looking at it, haven't thought about scripting. > We're interested in doing something like that > cliffw > > > > > > ------------------------------------------------------- > This SF.net email is sponsored by: Scholarships for Techies! > Can't afford IT training? All 2003 ictp students receive scholarships. > Get hands-on training in Microsoft, Cisco, Sun, Linux/UNIX, and more. > www.ictp.com/training/sourceforge.asp > _______________________________________________ > Lse-tech mailing list > Lse...@li... > https://lists.sourceforge.net/lists/listinfo/lse-tech |
From: <n0...@n0...> - 2003-01-22 20:46:14
|
If all you're looking for is regression issues have you thought about using `dbench' rather than `netbench'? `dbench' is a Linux benchmark that used traces from `netbench' to create the same traffic on a Linux samba server. As such, although I wouldn't guarantee it to create the exact same results as `netbench', it should be pretty comparable and it is certainly scriptable. On Wed, Jan 22, 2003 at 10:42:47AM -0800, Richard Sharpe wrote: > Hi, > > While there is an alternative to NetBench under development (or at least being talked about as part of SNIA and SPEC), there is also a need to be able to run NetBench. > > What I would like to do is to automate the running of NetBench and script it so that data collection can be driven from our existing UNIX-based infrastructure. This would be for checking for performance regressions etc. > > Has anyone looked at whether or not NetBench can be scripted in any way? > > Please reply to me directly, as I am not on the list. > > Regards > -- > Richard Sharpe > Software Engineer, Panasas Inc > Pioneering Smart and Infinitely Scalable Storage Networks > www.panasas.com, rs...@pa..., rs...@sa..., www.richardsharpe.com > > > ------------------------------------------------------- > This SF.net email is sponsored by: Scholarships for Techies! > Can't afford IT training? All 2003 ictp students receive scholarships. > Get hands-on training in Microsoft, Cisco, Sun, Linux/UNIX, and more. > www.ictp.com/training/sourceforge.asp > _______________________________________________ > Lse-tech mailing list > Lse...@li... > https://lists.sourceforge.net/lists/listinfo/lse-tech -- Don Dugger "Censeo Toto nos in Kansa esse decisse." - D. Gale n0...@ra... Ph: 303/652-0870x117 |
From: Richard S. <rs...@pa...> - 2003-01-22 21:28:46
|
On Wed, 22 Jan 2003 13:50:58 -0700 n0...@n0... wrote: > If all you're looking for is regression issues have you thought about > using `dbench' rather than `netbench'? `dbench' is a Linux benchmark > that used traces from `netbench' to create the same traffic on a > Linux samba server. As such, although I wouldn't guarantee it to create > the exact same results as `netbench', it should be pretty comparable and > it is certainly scriptable. We are interested in end-to-end performance, I guess, rather than just file system performance, and in any case, we already use SFS and other tools for filesystem performance. In addition, there is already the cifs_load_gen tool that takes pretty much the same approach as dbench/smbtorture. The issue is one of ensuring that 'reportable' numbers have not regressed. What I would really like in the short term is to callibrate cifs_load_gen against netbench and then use cifs_load_gen. > On Wed, Jan 22, 2003 at 10:42:47AM -0800, Richard Sharpe wrote: > > Hi, > > > > While there is an alternative to NetBench under development (or at least being talked about as part of SNIA and SPEC), there is also a need to be able to run NetBench. > > > > What I would like to do is to automate the running of NetBench and script it so that data collection can be driven from our existing UNIX-based infrastructure. This would be for checking for performance regressions etc. > > > > Has anyone looked at whether or not NetBench can be scripted in any way? > > > > Please reply to me directly, as I am not on the list. Regards -- Richard Sharpe Software Engineer, Panasas Inc Pioneering Smart and Infinitely Scalable Storage Networks www.panasas.com, rs...@pa..., sh...@sa..., www.richardsharpe.com |