After all this time, I found my fan turning on regularly and the same program
still.
I assume it reports package status back, fine with me.
BTW this is on landscape-client 24.02-0ubuntu5.3
But given what it is doing the cpu consumption to do so seems a bit much
for my taste and expectation.
sudo /usr/bin/time -v python3 /usr/bin/landscape-package-reporter
Command being timed: "python3 /usr/bin/landscape-package-reporter"
User time (seconds): 26.32
System time (seconds): 36.89
Percent of CPU this job got: 99%
Elapsed (wall clock) time (h:mm:ss or m:ss): 1:03.25
Average shared text size (kbytes): 0
Average unshared data size (kbytes): 0
Average stack size (kbytes): 0
Average total size (kbytes): 0
Maximum resident set size (kbytes): 229232
Average resident set size (kbytes): 0
Major (requiring I/O) page faults: 0
Minor (reclaiming a frame) page faults: 55168
Voluntary context switches: 10
Involuntary context switches: 3216
Swaps: 0
File system inputs: 0
File system outputs: 32
Socket messages sent: 0
Socket messages received: 0
Signals delivered: 0
Page size (bytes): 4096
Exit status: 0
AFAICS It does not spawn sub-processes - at least not in side the process
hierarchy.
So we'd want to profile the consumption of the script itself I guess?
Also that is quite some time in system, given that I assume it would be probing
the apt database right?
Checked behind which system calls that happens and got this:
% time seconds usecs/call calls errors syscall
------ ----------- ----------- --------- --------- ----------------
97.61 173.666872 3 48083857 39039084 newfstatat
1.48 2.630453 3 846972 fcntl
0.56 0.989142 3 273133 pread64
0.34 0.600204 2 212784 fstat
0.00 0.008634 2878 3 wait4
0.00 0.005327 4 1269 read
0.00 0.005153 1288 4 execve
0.00 0.005143 4 1057 munmap
0.00 0.004791 3 1329 mmap
0.00 0.003859 1286 3 clone
0.00 0.002790 3 832 49 openat
0.00 0.001694 2 793 close
0.00 0.001104 1 897 3 lseek
0.00 0.000876 4 179 getdents64
0.00 0.000623 1 457 448 ioctl
0.00 0.000422 3 134 brk
0.00 0.000248 3 63 mprotect
0.00 0.000194 2 91 rt_sigaction
0.00 0.000081 13 6 6 statfs
0.00 0.000049 9 5 write
0.00 0.000041 5 7 epoll_wait
0.00 0.000034 3 11 7 access
0.00 0.000026 8 3 sendto
0.00 0.000023 2 11 getrandom
0.00 0.000021 0 24 futex
0.00 0.000017 5 3 recvmsg
0.00 0.000014 1 13 epoll_ctl
0.00 0.000012 3 4 rseq
0.00 0.000011 2 4 getuid
0.00 0.000011 2 4 arch_prctl
0.00 0.000011 2 4 set_tid_address
0.00 0.000010 2 4 umask
0.00 0.000010 2 4 getgid
0.00 0.000009 1 7 set_robust_list
0.00 0.000009 2 4 prlimit64
0.00 0.000007 7 1 rt_sigreturn
0.00 0.000007 3 2 epoll_create1
0.00 0.000007 1 5 pipe2
0.00 0.000006 0 14 getpid
0.00 0.000004 2 2 msync
0.00 0.000004 0 10 9 readlink
0.00 0.000003 0 5 geteuid
0.00 0.000003 1 2 gettid
0.00 0.000002 1 2 getcwd
0.00 0.000000 0 9 dup2
0.00 0.000000 0 1 socket
0.00 0.000000 0 1 connect
0.00 0.000000 0 1 shutdown
0.00 0.000000 0 1 getsockopt
0.00 0.000000 0 3 uname
0.00 0.000000 0 1 flock
0.00 0.000000 0 1 getegid
0.00 0.000000 0 6 4 prctl
------ ----------- ----------- --------- --------- ----------------
100.00 177.927961 3 49424042 39039610 total
That is quite some unexpected fsstat and if that keeps my CPU busy for a
minute I wonder even more why :-/
I've looked a bit, but didn't have enough time.
Could be around _compute_packages_changes and its calling eventually causing
way too many find_index in apt_pkg.SourceList which takes 82ns each and happens
about a quarter million times. I feel this might be a red herring, so do not go
too much into that, but for completeness I'll attach a screenshot of a python
profile in kcachegrind.
Repro for me is as easy as:
$ sudo /usr/bin/time -v python3 /usr/bin/landscape-package-reporter
Maybe someone else could start cross checking if this is about a minute
of activity on every system or if a certain condition needs to be met.
** Attachment added: "kcachegrind (potential red herring though)"
https://bugs.launchpad.net/landscape-client/+bug/675577/+attachment/5883406/+files/Screenshot%20from%202025-06-11%2012-12-22.png
** Changed in: landscape-client (Ubuntu)
Status: Invalid => New
** Changed in: landscape-client
Status: Invalid => New
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/675577
Title:
landsape package reporter takes up too much cpu and memory resources
To manage notifications about this bug go to:
https://bugs.launchpad.net/landscape-client/+bug/675577/+subscriptions
--
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs