pcp
[Top] [All Lists]

[Bug 1325037] archive PM_MODE_FORW fetches not scanning to first occurre

To: pcp@xxxxxxxxxxx
Subject: [Bug 1325037] archive PM_MODE_FORW fetches not scanning to first occurrence of given metric
From: bugzilla@xxxxxxxxxx
Date: Fri, 08 Apr 2016 11:10:36 +0000
Auto-submitted: auto-generated
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <bug-1325037-355098@xxxxxxxxxxxxxxxxxxx>
References: <bug-1325037-355098@xxxxxxxxxxxxxxxxxxx>
https://bugzilla.redhat.com/show_bug.cgi?id=1325037



--- Comment #1 from Frank Ch. Eigler <fche@xxxxxxxxxx> ---
Created attachment 1145087
  --> https://bugzilla.redhat.com/attachment.cgi?id=1145087&action=edit
archive file (pmlogger against a container)

Possibly related; if a metric is completely absent, sometimes junk pmResult
comes back from a pmFetch.  With archive in newly attached tarball:

% PCP_DEBUG=65535  pmval -U FILE.meta hinv.ncpu

pmResult dump from 0x2577aa0 timestamp: 1456356755.639190 18:32:35.639 numpmid:
107
  60.26.0 (kernel.all.uptime): Bad file descriptor
  60.25.0 (kernel.all.nusers): Bad file descriptor
  60.7.67 (nfs4.server.reqs): Bad file descriptor
  60.7.66 (nfs4.server.calls): Bad file descriptor
  60.7.65 (nfs4.client.reqs): Bad file descriptor
  60.7.64 (nfs4.client.calls): Bad file descriptor
[...]
  [no hinv.ncpu metrics in this archive]

Those -EBADF codes are a mystery.  They also show up with pmdumplog.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
Unsubscribe from this bug 
https://bugzilla.redhat.com/token.cgi?t=Who9haI7y8&a=cc_unsubscribe
<Prev in Thread] Current Thread [Next in Thread>