ksyncer_daemon

[Stan PIetkiewicz_1]

rx7620 11.23 - when system gets busy, I notice the ksyncer_daemon process using large amounts of cpu cycles. It seems to be adding to the load of the server and can't seem to find out what it does or why it is so high in the processing queue. Have seen it using 180% (w%) of a cpu - something we never saw with the PA-RISC N class. This is a fairly highend 8way machine with 32gig of memory and don't understand why this process should be interfering with regular production. Any comments would be appreciated.

[spex]

Hi Colin,

I wasn't able to find any overt references to ksyncer_daemon. However, a little investigation reveals that it handles syncing for file system integrity.

From the following output, you can see that its ppid is 0 (swapper), and that it is started after lvmschedd but before lvmdevd.

# uname -mr
B.11.23 9000/800
# ps -ef | awk '{if ($3==0) print}' | sort -n -k2
root 0 0 0 Dec 8 ? 0:21 swapper
root 1 0 0 Dec 8 ? 0:09 init
root 2 0 0 Dec 8 ? 0:18 vhand
root 3 0 0 Dec 8 ? 0:48 statdaemon
root 4 0 0 Dec 8 ? 0:04 unhashdaemon
root 8 0 0 Dec 8 ? 0:00 kmemdaemon
root 9 0 0 Dec 8 ? 0:00 ioconfigd
root 10 0 0 Dec 8 ? 0:36 nfsktcpd
root 11 0 0 Dec 8 ? 0:00 autofskd
root 12 0 0 Dec 8 ? 0:00 lvmkd
root 13 0 0 Dec 8 ? 0:00 lvmkd
root 14 0 0 Dec 8 ? 0:00 lvmkd
root 15 0 0 Dec 8 ? 0:00 lvmkd
root 16 0 0 Dec 8 ? 0:00 lvmkd
root 17 0 0 Dec 8 ? 0:00 lvmkd
root 18 0 0 Dec 8 ? 0:00 lvmschedd
root 19 0 0 Dec 8 ? 1:04 ksyncer_daemon
root 20 0 0 Dec 8 ? 0:01 lvmdevd
root 21 0 0 Dec 8 ? 0:00 lvmattachd
root 22 0 0 Dec 8 ? 7:54 pagetableinit_daemon
root 23 0 0 Dec 8 ? 0:00 supsched
root 24 0 0 Dec 8 ? 0:00 strmem
root 25 0 0 Dec 8 ? 0:00 strweld
root 26 0 0 Dec 8 ? 0:00 strfreebd
root 27 0 0 Dec 8 ? 0:06 progressdaemon
root 28 0 0 Dec 8 ? 0:00 ttisr
root 29 0 0 Dec 8 ? 0:00 ipmid
root 36 0 0 Dec 8 ? 1:21 schedcpu
root 37 0 0 Dec 8 ? 0:06 pagezerod
root 38 0 0 Dec 8 ? 0:26 smpsched
root 39 0 0 Dec 8 ? 0:26 smpsched
root 40 0 0 Dec 8 ? 0:00 sblksched
root 41 0 0 Dec 8 ? 0:00 sblksched
root 42 0 0 Dec 8 ? 0:00 kend
root 48 0 0 Dec 8 ? 17:01 vxfsd
root 67 0 0 Dec 8 ? 0:00 rng

See syncer(1M) for more information.

PCS

[DCE]

Have you applied the latest synce patch? I believe it is PHKL_31500

[Stan PIetkiewicz_1]

Thanks for the comments ... will check the patches and let you know how it goes.

[Stan PIetkiewicz_1]

Hi ... did some checking and seems patches are uptodate. PHKL_31500 is there as well as PHKL_32055. Any suggestions? Seems that when there is a significant load on the server - high I/O as well, this process climbs to the top and consumes way more than we expect it to.

Thanks.