|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 24.02.16 3:38, Heiler Bemerguy пишет: > > > 23/02/2016 16:40, Yuri Voinov wrote: >> >> When you CPU's/cores waiting for HDD access, they got high-loag. >> > > Are you sure it would show up as "User" load and not as "Wait" ? Depending OS. Most likely "User" exactly. Still looks like overloaded IO queue. > On linux "TOP" it shows something like: > %Cpu0 : 99,0 *us*, 1,0 sy, 0,0 ni, 0,0 id, 0,0 *wa*, 0,0 hi, 0,0 si, 0,0 st top is not performance tool. It hides most important aspects of performance-related problem. Use it only at a glance tool. For performance tuning usually uses more specific (and OS-specific, of course) tools. iostat, vmstat, swap -l, sar etc. > > > 23/02/2016 16:55, Amos Jeffries wrote: >> What GHz rating is each CPU core? 200-250 RPS is roughly in the range I >> would expect from a 1.xGHz core going full speed / 100% usage. >> >> Are you using RAID on the disk storage? IME, RAID can more than halve >> the speed of the proxy. Although the CPU thrashing effect is mostly >> hidden away out of sight in the disk controller processor(s). > > 6-core of Intel(R) Xeon(R) CPU E5-2667 v2 @ 3.30GHz > It's a vmware VM with a NetApp 266TB storage. This VM has access to 2 different LUNs in 2 different controllers to a bunch SAS HDs, > dd if=/dev/zero of=testfile bs=2G count=1 oflag=dsync = ~391 MB/s This is tells nothing about SAN topology and whole IO subsystem. There is much questions occurs here: - - Which connection uses HDD's bunches? FC? iSCSI? FCoE? If iSCSI/some_ethernet-Based - did you turned on jumbograms? - - How much hardware cache installed in controllers and which type? - - How was created LUN's? - - Which RAID type you uses and what is it parameters? - - How configured VMWare? Does you configure deferred syncing VM caches? Or not? etc.etc.etc. Note: NetApp (you have forgotten to specify your NetApp model and licensed options) has known issues with: CoW with WAFL, hardware read-cache and write-cache - this is product specific options sold separately (!) and, generally, tuning NetApp is separated complex task. Especially with virtualized environment. And please note: One DD's file - it's about nothing. This is unrelated to REAL load. As by as abstract IOPS numbers. > > Best Regards, > -- > Heiler Bemerguy - (91) 98151-4894 > Assessor Técnico - CINBESA (91) 3184-1751 > > > _______________________________________________ > squid-users mailing list > squid-users@xxxxxxxxxxxxxxxxxxxxx > http://lists.squid-cache.org/listinfo/squid-users -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQEcBAEBCAAGBQJWzfdpAAoJENNXIZxhPexGYnwIALmxL3FA75PfEiaeHeRXu4a5 kFuzHgSb9pjldlOPHxorgbVul3BU5HE+LVFmzYsD0lMIK9hbqutKz/m/ji5B0QBN yDevoIi+rGKUgXTGs0VvPZprLzbCc9X1V7uIQeCFne1gvWruyr4dx4xrjZSONdUP 8wcNN5irNZ0Zn+fcqQUK+G9vmvHWDBeQK/gjOFk4ZdUsWZkNx4vB7RGeBBTSH5Lo G/vzY4j/Cd0L89B1PPrWrrTdko0jGe4Z0D4kFaA+bNb+eV56uRx0m7Qii4Cnl4Vu KGSHhwD6gaI0sJGwWqwRgcNEPEtMcAGtO8IvaWmHGVGCEz+Bf5NyAbAtF7cQ9Zc= =zPnN -----END PGP SIGNATURE----- |
Attachment:
0x613DEC46.asc
Description: application/pgp-keys
_______________________________________________ squid-users mailing list squid-users@xxxxxxxxxxxxxxxxxxxxx http://lists.squid-cache.org/listinfo/squid-users
