Re: possible leak in kernel 2.6.10-ac12

2005-02-19 Thread Parag Warudkar
On Saturday 19 February 2005 07:37 am, Pedro Venda wrote: > biovec-1          1989252 1989478     16  226    1 : tunables  120   60   >  0 : slabdata   8803   8803      0 bio               1989270 1989271     64 >   61    1 : tunables  120   60    0 : slabdata  32611  32611      0 You have bio lea

Re: possible leak in kernel 2.6.10-ac12

2005-02-19 Thread Pedro Venda
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Parag Warudkar wrote: | On Wednesday 16 February 2005 06:28 pm, Pedro Venda wrote: | |>Having upgraded most of them to 2.6.10-ac12, one of them showed a linear |>growth of used memory over the last 7 days, after the first 2.6.10-ac12 |>boot. It came to

Re: possible leak in kernel 2.6.10-ac12

2005-02-17 Thread Bill Davidsen
Parag Warudkar wrote: On Wednesday 16 February 2005 06:28 pm, Pedro Venda wrote: Having upgraded most of them to 2.6.10-ac12, one of them showed a linear growth of used memory over the last 7 days, after the first 2.6.10-ac12 boot. It came to a point that it started swapping and the swap usage too

Re: possible leak in kernel 2.6.10-ac12

2005-02-17 Thread Maciej Soltysiak
Hello Pedro, Thursday, February 17, 2005, 12:28:15 AM, you wrote: > boot. It came to a point that it started swapping and the swap usage too > started to grow linearly. I had the same with swap being eaten especially by perl apps like qmail-scanner I think this helps: --- a/mm/vmscan.c 200

Re: possible leak in kernel 2.6.10-ac12

2005-02-16 Thread Pedro Venda
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Pedro Venda wrote: | Noel Maddy wrote: | | On Thu, Feb 17, 2005 at 12:02:33AM +, Pedro Venda wrote: | | | | | |>admin proc # cat slabinfo | | | | ... | | | |>biovec-1 74224 74354 16 2261 : tunables 120 | 600 : | slabdata

Re: possible leak in kernel 2.6.10-ac12

2005-02-16 Thread Pedro Venda
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Noel Maddy wrote: | On Thu, Feb 17, 2005 at 12:02:33AM +, Pedro Venda wrote: | | |>admin proc # cat slabinfo | | ... | |>biovec-1 74224 74354 16 2261 : tunables 120 600 : slabdata329329 0 |>bio

Re: possible leak in kernel 2.6.10-ac12

2005-02-16 Thread Noel Maddy
On Thu, Feb 17, 2005 at 12:02:33AM +, Pedro Venda wrote: > admin proc # cat slabinfo ... > biovec-1 74224 74354 16 2261 : tunables 120 600 : > slabdata329329 0 > bio74212 74237 64 611 : tunables 120 600 : > slabdata

Re: possible leak in kernel 2.6.10-ac12

2005-02-16 Thread Pedro Venda
Parag Warudkar wrote: > On Wednesday 16 February 2005 06:28 pm, Pedro Venda wrote: > >>Having upgraded most of them to 2.6.10-ac12, one of them showed a linear >>growth of used memory over the last 7 days, after the first 2.6.10-ac12 >>boot. It came to a point that it started swapping and the swap

Re: possible leak in kernel 2.6.10-ac12

2005-02-16 Thread Parag Warudkar
On Wednesday 16 February 2005 06:28 pm, Pedro Venda wrote: > Having upgraded most of them to 2.6.10-ac12, one of them showed a linear > growth of used memory over the last 7 days, after the first 2.6.10-ac12 > boot. It came to a point that it started swapping and the swap usage too > started to gro

possible leak in kernel 2.6.10-ac12

2005-02-16 Thread Pedro Venda
hi everyone, I maintain a couple of linux servers in an academic environment. Having upgraded most of them to 2.6.10-ac12, one of them showed a linear growth of used memory over the last 7 days, after the first 2.6.10-ac12 boot. It came to a point that it started swapping and the swap usage too st