| List: | General Discussion | « Previous MessageNext Message » | |
| From: | Fred Lindberg | Date: | June 28 1999 7:40pm |
| Subject: | Re: to much fragmentation | ||
| View as plain text | |||
On Mon, 28 Jun 1999 10:40:05 -0700, Scott Hess wrote: >What we've found to happen is that all of your clients are hitting the >database, and at some point it catastrophically degrades - performance >doesn't plateau (sp?), it degrades. Our best simulations indicate that this >point is where the number of transactions per second to the disk drive >passes a certain point (this is as opposed to megabytes per second). At This is a problem of your OS. Modern OS's, e.g. linux, will use unused memory as cache. Performance will degrade when you don't have memory to hold the blobs. In this case, you don't have memory to hold the blob table and thus access to the db would require disk access. At that point it depends on which system is better at caching the blobs. (AFAIK) -Sincerely, Fred (Frederik Lindberg, Infectious Diseases, WashU, St. Louis, MO, USA)
| Thread | ||
|---|---|---|
| • to much fragmentation | David Johnson | 26 Jun |
| • Re: to much fragmentation | Sasha Pachev | 27 Jun |
| • Re: to much fragmentation | Scott Hess | 28 Jun |
| • Re: to much fragmentation | Vivek Khera | 28 Jun |
| • Re: to much fragmentation | Michael Widenius | 29 Jun |
| • Re: to much fragmentation | Ed Carp | 29 Jun |
| • Re: to much fragmentation | Sasha Pachev | 29 Jun |
| • Re: to much fragmentation | Michael Widenius | 29 Jun |
| • Authentication methods | Roger Smith | 29 Jun |
| • Authentication methods | Jani Tolonen | 1 Jul |
| • Re: to much fragmentation | Scott Hess | 29 Jun |
| • Re: to much fragmentation | Ed Carp | 29 Jun |
| • Re: to much fragmentation | Michael Widenius | 30 Jun |
| • Re: to much fragmentation | Tõnu Samuel | 29 Jun |
| • Re: to much fragmentation | Scott Hess | 28 Jun |
| • Re: to much fragmentation | Fred Lindberg | 28 Jun |
