I realized now that this measurement is wrong. "vmstat -iz" seems to calculate the interrupt rate based a longer period, and this measurement was taken just after we started to push traffic through the machine again.
The amount of interrupts per second when checking with systat (which seems to have a shorter measurement period) at the same time was way higher, about 5000 intr/s on em0 and em2. Sorry for the wrong data On 2011-03-28 17:46, Peter Hallin wrote: > This is the output when the machine is running at 80 Mbit/s > and CPU usage is almost 100% interrupts: > > Please note that this is after we rebooted with the SP kernel, > which didn't make any differences. > > systat ifs: > > IFACE STATE DESC IPKTS IBYTES IERRS OPKTS > OBYTES OERRS COLLS > em0 up:U 6772 8285095 0 5043 > 1251821 0 0 > em1 dn:D 0 0 0 0 > 0 0 0 > em2 up:U 5050 1268574 0 6762 > 8280262 0 0 > em3 dn:D 0 0 0 0 > 0 0 0 > em4 up:U 19 7136 0 774 > 361639 0 > > vmstat -iz: > > interrupt total rate > irq0/clock 73974 100 > irq144/acpi0 0 0 > irq112/ppb2 0 0 > irq112/em0 767598 1038 > irq113/em1 0 0 > irq113/ppb3 0 0 > irq113/em2 1007863 1363 > irq114/em3 0 0 > irq114/em4 460696 623 > irq115/em5 0 0 > irq112/ppb7 0 0 > irq96/uhci0 0 0 > irq97/uhci1 0 0 > irq96/uhci2 0 0 > irq97/uhci3 0 0 > irq96/ehci0 0 0 > irq98/pciide0 2083 2 > irq99/pciide1 0 0 > irq99/ichiic0 0 0 > irq145/com0 0 0 > irq146/com1 0 0 > irq147/pckbc0 304 0 > irq148/pckbc0 0 0 > Total 2312518 3129 > > > The next step for us will be to swap the PCIe NICs to > regular PCI to see if that does anything. > > > On 2011-03-28 13:55, Jonathan Gray wrote: > > It sounds like you have an interrupt storm what does > > vmstat -iz show? > -- -- Peter Hallin | Tel: +46 46 2221341 IT-Sdkerhet och brandvdggar | Fax: +46 46 2221399 LDC, Lunds Universitet | E-post: peter.hal...@ldc.lu.se Margaretav. 1A, 222 40, LUND | http://www.ldc.lu.se/security/
