Note: This bug is displayed in read-only format because
the product is no longer active in Red Hat Bugzilla.
RHEL Engineering is moving the tracking of its product development work on RHEL 6 through RHEL 9 to Red Hat Jira (issues.redhat.com). If you're a Red Hat customer, please continue to file support cases via the Red Hat customer portal. If you're not, please head to the "RHEL project" in Red Hat Jira and file new tickets here. Individual Bugzilla bugs in the statuses "NEW", "ASSIGNED", and "POST" are being migrated throughout September 2023. Bugs of Red Hat partners with an assigned Engineering Partner Manager (EPM) are migrated in late September as per pre-agreed dates. Bugs against components "kernel", "kernel-rt", and "kpatch" are only migrated if still in "NEW" or "ASSIGNED". If you cannot log in to RH Jira, please consult article #7032570. That failing, please send an e-mail to the RH Jira admins at rh-issues@redhat.com to troubleshoot your issue as a user management inquiry. The email creates a ServiceNow ticket with Red Hat. Individual Bugzilla bugs that are migrated will be moved to status "CLOSED", resolution "MIGRATED", and set with "MigratedToJIRA" in "Keywords". The link to the successor Jira issue will be found under "Links", have a little "two-footprint" icon next to it, and direct you to the "RHEL project" in Red Hat Jira (issue links are of type "https://issues.redhat.com/browse/RHEL-XXXX", where "X" is a digit). This same link will be available in a blue banner at the top of the page informing you that that bug has been migrated.
DescriptionCharles Haithcock
2021-07-23 19:19:19 UTC
Description of problem:
PCP by default monitors /proc/zoneinfo and collects everything there. However, DMA and DMA32 zones will never have memory in them beyond Node 0. These zones are (mostly) dedicated to device IO and must reside on the lowest addresses in NUMA Node 0, as devices can use 8 and 16 bit addressing. Similarly, the kernel hardcodes the upper limits of physical memory address ranges for those two zone types. The kernel simply lazily enumerates them for other NUMA nodes as well for code efficiency but does not put memory into them.
PCP should not monitor DMA or DMA32 for NUMA Nodes above Node 0 as it will never be useful. Doing so will help reduce the amount of data collected and should see a small reduction in pmlogger sizes (which will be more noticable on considerably larger systems with many many NUMA nodes, this will be a negligible change for systems with much fewer Nodes).
Version-Release number of selected component (if applicable):
$ grep -e ^pcp -e ^pm installed-rpms
pcp-5.2.5-4.el8.x86_64 Tue Jun 15 22:23:52 2021
pcp-conf-5.2.5-4.el8.x86_64 Tue Jun 15 22:23:51 2021
pcp-libs-5.2.5-4.el8.x86_64 Tue Jun 15 22:23:51 2021
How reproducible:
100%
Steps to Reproduce:
1.
2.
3.
Actual results:
Tracking guaranteed 0s for DMA and DMA32 zones for NUMA nodes above node 0
Expected results:
Those metrics are never tracked to begin with
Additional info:
From the customer's pcp data (as an example):
$ pminfo -tf -a 20210519.0.xz mem.zoneinfo.spanned
mem.zoneinfo.spanned [spanned space in each zone for each NUMA node]
inst [0 or "DMA::node0"] value 16380
inst [1 or "Normal::node0"] value 99221504
inst [2 or "Normal::node1"] value 102760448
inst [3 or "Device::node1"] value 0
inst [4 or "DMA32::node0"] value 4177920
inst [5 or "Movable::node0"] value 0
inst [6 or "Device::node0"] value 0
inst [7 or "DMA::node1"] value 0 <---
inst [8 or "DMA32::node1"] value 0 <---
inst [9 or "Movable::node1"] value 0
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory (pcp bug fix and enhancement update), and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.
https://access.redhat.com/errata/RHBA-2022:1765
Description of problem: PCP by default monitors /proc/zoneinfo and collects everything there. However, DMA and DMA32 zones will never have memory in them beyond Node 0. These zones are (mostly) dedicated to device IO and must reside on the lowest addresses in NUMA Node 0, as devices can use 8 and 16 bit addressing. Similarly, the kernel hardcodes the upper limits of physical memory address ranges for those two zone types. The kernel simply lazily enumerates them for other NUMA nodes as well for code efficiency but does not put memory into them. PCP should not monitor DMA or DMA32 for NUMA Nodes above Node 0 as it will never be useful. Doing so will help reduce the amount of data collected and should see a small reduction in pmlogger sizes (which will be more noticable on considerably larger systems with many many NUMA nodes, this will be a negligible change for systems with much fewer Nodes). Version-Release number of selected component (if applicable): $ grep -e ^pcp -e ^pm installed-rpms pcp-5.2.5-4.el8.x86_64 Tue Jun 15 22:23:52 2021 pcp-conf-5.2.5-4.el8.x86_64 Tue Jun 15 22:23:51 2021 pcp-libs-5.2.5-4.el8.x86_64 Tue Jun 15 22:23:51 2021 How reproducible: 100% Steps to Reproduce: 1. 2. 3. Actual results: Tracking guaranteed 0s for DMA and DMA32 zones for NUMA nodes above node 0 Expected results: Those metrics are never tracked to begin with Additional info: From the customer's pcp data (as an example): $ pminfo -tf -a 20210519.0.xz mem.zoneinfo.spanned mem.zoneinfo.spanned [spanned space in each zone for each NUMA node] inst [0 or "DMA::node0"] value 16380 inst [1 or "Normal::node0"] value 99221504 inst [2 or "Normal::node1"] value 102760448 inst [3 or "Device::node1"] value 0 inst [4 or "DMA32::node0"] value 4177920 inst [5 or "Movable::node0"] value 0 inst [6 or "Device::node0"] value 0 inst [7 or "DMA::node1"] value 0 <--- inst [8 or "DMA32::node1"] value 0 <--- inst [9 or "Movable::node1"] value 0