Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[DNM] jewel: mon: force_create_pg could leave pg stuck in creating state #13789

Closed
wants to merge 20 commits into from
Closed

[DNM] jewel: mon: force_create_pg could leave pg stuck in creating state #13789

wants to merge 20 commits into from

Conversation

shinobu-x
Copy link
Contributor

@shinobu-x shinobu-x commented Mar 3, 2017 •
edited

@shinobu-x shinobu-x changed the title jewel: mon: force_create_pg could leave pg stuck in creating state DNM: jewel: mon: force_create_pg could leave pg stuck in creating state Mar 3, 2017
@liewegas @shinobu-x
mon/MonCommands: remove send_pg_creates from cmd table
25dba7e 
Forgot this in 6cbdd67

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 15f17bc)
(cherry picked from commit 998d990)
@shinobu-x shinobu-x force-pushed the wip-18298-jewel branch 2 times, most recently from 28b405d to 25dba7e Compare March 4, 2017 04:43
@liewegas @shinobu-x
mon/PGMonitor: clean up some send_pg_creates cruft
ca252fd 
ff1fd4b already removed most of
this; remove the now useless stub and clean up the unneede con
check.

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit b79a85e)

Conflicts:
	src/mon/PGMonitor.cc
@shinobu-x
Copy link
Contributor Author

jenkins test this please

@shinobu-x
Adding missing void send_pg_creates() in src/mon/PGMonitor.h.
8d8e527 
Signed-off-by: Shinobu Kinjo <shinobu@redhat.com>
@shinobu-x shinobu-x force-pushed the wip-18298-jewel branch 4 times, most recently from 09c4e65 to 8d8e527 Compare March 5, 2017 00:10
@xiexingguo @shinobu-x
mon/PGMonitor: simplify propose logic for check_osd_map() process
f6f1a95 
We'll assure pg_map.last_osdmap_epoch is smaller than osdmap's epoch
at the function entry, so we are doomed to update pending_inc.osdmap_epoch,
which means the "propose" variables will be always set to true.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
(cherry picked from commit ea1cb9f)
@liewegas @shinobu-x
include/intarith: add ctz/clz/cbits helpers
50cfdaf 
"Count trailing zero" and "count leading zero".

Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 04a3ce6)
@shinobu-x shinobu-x force-pushed the wip-18298-jewel branch 5 times, most recently from fcc2c2a to 50cfdaf Compare March 6, 2017 00:07
hjwsm1989 and others added 9 commits March 6, 2017 15:17
@hjwsm1989 @shinobu-x
mon: print the num_pools and num_objects in 'ceph -s -f json/json-pr…
621d185 
…etty'.

  Fixes: http://tracker.ceph.com/issues/17703
  Signed-off-by: huangjun <hjwsm1989@gmail.com>

(cherry picked from commit bf2b50e)
@xiexingguo @shinobu-x
mon/PGMap: use TextTable for dump_pg_sum_stats()
a092eb0 
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
(cherry picked from commit 2898b7b)
@xiexingguo @shinobu-x
mon/PGMAP: use TextTable for dump_osd_stats()
55debc4 
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
(cherry picked from commit 732f2da)

Conflicts:
	src/mon/PGMap.cc
@xiexingguo @shinobu-x
mon/PGMAP: use TextTable for dump_osd_sum_stats()
9789fa9 
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
(cherry picked from commit 77db9d8)

Conflicts:
	src/mon/PGMap.cc
@xiexingguo @shinobu-x
mon/PGMap: use TextTable for dump_filtered_pg_stats()
3892c90 
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
(cherry picked from commit fc33cf4)
@xiexingguo @shinobu-x
mon/PGMonitor: batch filter pg state for "pg ls *" command family
184c032 
Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
(cherry picked from commit fbb19fc)

Conflicts:
	src/mon/PGMonitor.cc
@yonghengdexin735 @shinobu-x
mon: remove the redundant min_last_epoch_clean judge in PGMap
e9ff8ba 
 In the end of the apply_incremental function, no matter what

 happens the min_last_epoch_clean must be 0.

 Signed-off-by:song baisen <song.baisen@zte.com.cn>

(cherry picked from commit 7b03828)
@xiexingguo @shinobu-x
mon/PGMap: fix wrong size of dump_* commands
dcfcc45 
The prettybyte_t considers the input as bytes, so the
pre-cast is needed here.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
(cherry picked from commit f658c8d)
@xiexingguo @shinobu-x
mon/PGMap: use si_t for space translation
f00f507 
which is simpler and better for alignment.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
(cherry picked from commit c3c62e8)
@xiexingguo @shinobu-x
mon/PGMap: make 1st column of output of textable left justified
e912c1c 
So the whole output looks gorgeous.

Signed-off-by: xie xingguo <xie.xingguo@zte.com.cn>
(cherry picked from commit bf9f5c1)

Conflicts:
	src/mon/PGMap.cc
@shinobu-x
common/TextTable: add include guards
46b67da 
Signed-off-by: John Spray <john.spray@redhat.com>
(cherry picked from commit ab24038)
@shinobu-x
Copy link
Contributor Author

retest this please

@shinobu-x
mon: move "df" dump code from PGMonitor to PGMap
318a9f3 
...so that ceph-mgr can use it too.

Signed-off-by: John Spray <john.spray@redhat.com>
(cherry picked from commit 519a01d)

Conflicts:
	src/mon/PGMap.cc
	src/mon/PGMonitor.cc
@shinobu-x
Copy link
Contributor Author

retest this please

@shinobu-x
mon: refactor PGMap updating code for reuse in mgr
332937b 
...and remove the copypasta from mgr.

mgr still doesn't do all the same logic (it doesn't
have osdmap history handy) but it is now much easier
to see which bits are used one place vs. the other.

Signed-off-by: John Spray <john.spray@redhat.com>
(cherry picked from commit afa7078)

Conflicts:
	src/mgr/ClusterState.cc
	src/mon/PGMap.h
	src/mon/PGMonitor.cc
	src/osd/OSDMap.h
@shinobu-x shinobu-x force-pushed the wip-18298-jewel branch 2 times, most recently from 6cbb460 to f3d18ac Compare March 8, 2017 04:52
@shinobu-x shinobu-x force-pushed the wip-18298-jewel branch 3 times, most recently from e2ccb01 to f3d18ac Compare March 9, 2017 07:41
@shinobu-x
mon: remove need for friending OSDMap to PGMapUpdater
b2d27f0 
This used to be the friend relationship to PGMonitor.  It
was never really needed anyway, it was only being used
to skip the getter on `pools`

Signed-off-by: John Spray <john.spray@redhat.com>
(cherry picked from commit 78b6eeb)

Conflicts:
	src/osd/OSDMap.h
@liewegas @shinobu-x
mon/PGMonitor: fix force_create_pg
c315c5e 
We weren't carefully registering the creating PG.  In
particular, the current osd mappings (acting, up, etc.)
weren't getting filled in, which meant the PG wasn't
(necessarily) mapped to an OSD until the OSDMap updated
and we recalculated mappings.

Fix by sending us through the common code path so that
all fields get updated properly.

Fixes: http://tracker.ceph.com/issues/18298
Signed-off-by: Sage Weil <sage@redhat.com>
(cherry picked from commit 9a41a0b)

Conflicts:
	src/mon/PGMonitor.cc
@shinobu-x shinobu-x changed the title DNM: jewel: mon: force_create_pg could leave pg stuck in creating state jewel: mon: force_create_pg could leave pg stuck in creating state Mar 9, 2017
@shinobu-x
Copy link
Contributor Author

retest this please

@smithfarm
Copy link
Contributor

@jdurgin This is a backport of #12539 . . . at first glance, it appears to be backporting too much.

@smithfarm smithfarm requested a review from jdurgin April 13, 2017 14:46
@smithfarm smithfarm added this to the kraken milestone Apr 13, 2017
@smithfarm smithfarm changed the title jewel: mon: force_create_pg could leave pg stuck in creating state [DNM] jewel: mon: force_create_pg could leave pg stuck in creating state Apr 13, 2017
@jdurgin
Copy link
Member

jdurgin commented Apr 28, 2017

@smithfarm I agree, this is backporting too much. We really just want the minimal way to fix this issue, even if it's not a clean cherry-pick and needs some extra editing for jewel.

@smithfarm smithfarm closed this Jul 7, 2017
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@jdurgin jdurgin Awaiting requested review from jdurgin

Assignees
No one assigned
Labels
bug-fix
Projects
None yet
Milestone
kraken
7 participants
@shinobu-x @smithfarm @jdurgin @liewegas @xiexingguo @hjwsm1989 @yonghengdexin735