Hi Heikki and Tom,
>I bet it's the GCC compile farm's storage system that sucks. Buildfarm
>animal 'douc' doesn't seem to have that problem:
We have setup the buildframs on this node in OSU lab for community usage.
Let me know if you want to run your testing on this node (douc).
I can provide you with the access.
I see there are multiple plog-con running on this node. This done by the make
community and they are using the same filesystems that you are using. It would
be better to use a different node for performance measurement.
Topas Monitor for host:gcc119 EVENTS/QUEUES FILE/TTY
Wed Feb 4 06:43:28 2026 Interval:2 Cswitch 470 Readch 298 Syscall 3381.3K Writech 19189
CPU User% Kern% Wait% Idle% Physc Entc% Reads 3379.3K Rawin 0
Total 24.9 33.5 0.0 41.6 4.02 100.47 Writes 979 Ttyout 298 Forks 0 Igets 7
Network BPS I-Pkts O-Pkts B-In B-Out Execs 0 Namei 534
Total 1.76K 31.50 0.50 1.42K 351.0 Runqueue 5.50 Dirblk 0 Waitqueue 0.0
Disk Busy% BPS TPS B-Read B-Writ MEMORY
Total 7.9 626K 156.5 0 626K PAGING Real,MB 163840 Faults 0 % Comp 16
FileSystem BPS TPS B-Read B-Writ Steals 0 % Noncomp 36
Total 18.4K 133.5 298.5 18.1K PgspIn 0 % Client 36 PgspOut 0
Name PID CPU% PgSp Owner PageIn 0 PAGING SPACE
plog-con 35520802 16.3 304K bradking PageOut 0 Size,MB 2048
plog-con 29753802 15.3 308K bradking Sios 0 % Used 5
plog-con 31392046 15.3 308K bradking % Free 95
plog-con 35914140 14.3 304K bradking NFS (calls/sec)
tar 23789996 0.2 280K sortie SerV2 0 WPAR Activ 0
topas 19661074 0.0 9.48M root CliV2 0 WPAR Total 0
getty 18415934 0.0 664K root SerV3 0 Press: "h"-help
rpc.lock 8126804 0.0 1.19M root CliV3 0 "q"-quit
Warm regards,
Sriram.
