Ceph Cheat Sheet

by Eagle1992 (Eagle1992) via cheatography.com/195419/cs/40944/

ceph config

ceph config dump Dumps entire monitor config​uration database

ceph config get {who} {option} Dumps the config​uration option stored in monitor config​uration database for specific daemon or client(for
example osd.0). Options können angegeben werden.

ceph config set {who} {option} specifies a config​uration option in the monitor config​uration database (for example ceph config set osd.0
{value} debug_ms 20)
ceph config show {who} shows runtime settings for a running daemon (to see all settings use: ceph config show-w​ith​-de​faults)
ceph config assimi​lat​e-conf -i ingests a config​uration file from inpute file and moves any valid options into the monitor config​uration
{input​_file} -o {outpu​t_file} database
ceph config help {option} (-f json-p​‐ to get help for particular option (not optional).
retty)

ceph tell {who} config set {option} set tempor​arily other settings (for example ceph tell osd.123 config set debug_osd 20) You can also
{value} specify wildcards: osd.* (to change settings for all OSDs.)

ceph is part of Ceph, a massively scalable, open-s​ource, distri​buted storage system. Please refer to the Ceph docume​ntation at https:​//d​ocs.ce​‐
ph.com for more inform​ation.

Options Options (cont) Options (cont)

-i infile will specify an input file to be

passed along as a payload with
the command to the monitor
cluster. This is only used for
specific monitor commands.
-o outfile will write any payload returned
by the monitor cluster with its
reply to outfile. Only specific
monitor commands (e.g. osd
getmap) return a payload.
-- will apply the approp​riate user
setuser ownership to the file specified by
user the option ‘-o’.
-- will apply the approp​riate group
setgroup ownership to the file specified by
group the option ‘-o’.
 -c ceph.conf, Use ceph.conf config​‐ --watc​h- Watch warning events.
--conf​=ce​‐ uration file instead of the warn
ph.conf default /etc/c​eph​/ce​ph.conf --watc​h- Watch error events.
to determine monitor error
addresses during startup.
-- Display version.
--id Client id for authen​tic​ation. version,
CLIENT_ID, -v
--user
-- Make verbose.
CLIENT_ID
verbose
--name Client name for authen​tic​‐
-- Make less verbose.
CLIENT​‐ ation.
concise
_NAME, -n
-f {json,​‐ Format of output. Note: yaml is
CLIENT​‐
jso​n-p​‐ only valid for orch commands.
_NAME
ret​ty,​‐
--cluster Name of the Ceph cluster.
xml​,xm​l-
CLUSTER
p​ret​ty,​‐
--admi​n-d​‐ Submit admin-​socket pla​in,​‐
aemon commands via admin yaml}, --
ADMIN_​‐ sockets in /var/r​un/​ceph. format
SOCKET,
--conn​‐ Set a timeout for connecting to
daemon
ect​-ti​‐ the cluster.
DAEMON​‐
meout
_NAME
CLUSTE​
-s, --status Show cluster status. R_T​‐
-w, --watch Watch live cluster changes IMEOUT
on the default ‘cluster’ --no-i​‐ --no-i​ncr​easing is off by default.
channel ncr​‐ So increasing the osd weight is
-W, --watc​h- Watch live cluster changes easing allowed using the reweig​ht-​by-​‐
c​hannel on any channel (cluster, uti​liz​ation or test-r​ewe​igh​t-b​y-u​‐
audit, cephadm, or * for all) til​ization commands. If this
--watc​h- Watch debug events. option is used with these
debug commands, it will help not to
increase osd weight even the
--watc​h-info Watch info events.
osd is under utilized.
--watc​h-sec Watch security events.
--block block until completion (scrub
and deep-scrub only)

By Eagle1992 (Eagle1992) Published 23rd October, 2023. Sponsored by CrosswordCheats.com

cheatography.com/eagle1992/ Last updated 23rd October, 2023. Learn to solve cryptic crosswords!
Page 2 of 8. http://crosswordcheats.com
 Ceph Cheat Sheet
by Eagle1992 (Eagle1992) via cheatography.com/195419/cs/40944/

ceph mon ceph mgr (cont) Miscel​laneous (cont)

ceph mon dumps formatted monmap - if ceph mgr will report metadata about all ceph balancer enable / disable ceph
dump integer is given, then you'll get metadata manager daemons or if the off / on balancer
{int0-} mon-map from epoch {integer] {name} name is specified a single ceph balancer set balancer mode to
ceph mon adds new monitor named manager daemon mode crush-​‐ crush-​compat or upmap
add {name} at {addr} ceph mgr will report a count of running compat or (default)
{name} versions daemon versions upmap
{IPadd​r[:​‐ ceph mgr will report a count of any ceph -s / -- see actual ceph status
port]} count-​‐ daemon metadata field status
ceph mon gets monmap (from specified met​adata ceph df detail shows data usage in raw
getmap epoch) {field} storage and pools
{int0-}
ceph-v​olume shows all disks
ceph mon removes monitor named Miscel​laneous
lvm list or
remove {name} ceph tell mon.<i​d> Cause a specific ceph device ls
{name} quorum enter|exit MON to enter or exit
ceph crash ls shows all mgr module
ceph mon summarizes monitor status quorum.
/ ls-new crash dumps (or only list
stat ceph quorum​‐ Reports status of new crash dumps with ls-
_status monitor quorum. new)
ceph mgr
ceph report {<t​ags> Reports full status of ceph crash shows exact inform​ation
ceph mgr dumps latest MgrMap, which [<t​ags​>...]} cluster, optional title info {crashid} for crashdump with
dump describes the active & standby tag strings. specific crashid
manager daemons ceph status Shows cluster ceph crash archive all crash dumps
ceph mgr will mark a manager daemon status. archiv​e-all
fail as failed, removing it from ceph tell <name Sends a command ceph crash rm removes crash dump with
{name} managerr map (type.i​d)> <co​mma​‐ to a specific {crashid} specific id
ceph mgr will list currently enableld nd> [optio​ns...] daemon.
ceph crash List the timest​amp​/uuid
module ls manager modules (plugins) ceph tell <name List all available stat crashids for all newcrash
ceph mgr will enable a manager modules. (type.i​d)> help commands. info.
module available modules are included ceph version Show mon daemon ceph crash Remove saved crashes
enable in MgrMap and visible via mgr version prune {keep} older than ‘keep’ days.
{module} dump
ceph fs dump get MDS Map {keep} must be an integer.
ceph mgr will disable an active manager
module module
ceph balancer get status of ceph
disable
status balancer
{module}

By Eagle1992 (Eagle1992) Published 23rd October, 2023. Sponsored by CrosswordCheats.com

cheatography.com/eagle1992/ Last updated 23rd October, 2023. Learn to solve cryptic crosswords!
Page 3 of 8. http://crosswordcheats.com
 Ceph Cheat Sheet
by Eagle1992 (Eagle1992) via cheatography.com/195419/cs/40944/

Miscel​laneous (cont) ceph pg (cont) ceph pg (cont)

ceph crash Archive a crash report so that ceph pg dump {all|s​umm​‐ shows ceph pg repair starts repair on <pg​id>.
archive it is no longer considered for ary​|su​m|d​elt​a|p​ool​s|o​‐ human-​rea​‐ <pg​id>
{crashid} the RECENT​_CRASH health sds​|pg​s|p​gs_​brief} [{all|​‐ dable ceph pg scrub starts scrub on <pg​id>.
check and does not appear in sum​mar​y|s​um|​del​ta|​poo​‐ versions of pg <pg​id>
the crash ls-new output (it will ls|​osd​s|p​gs|​pgs​_br​ief...]} map (only ‘all’
ceph pg stat shows placement group
still appear in the crash ls valid with
status.
output). plain).
https:​//d​ocs.ce​ph.c​om​/en​/qu​inc​y/r​ado​s/o​‐
ceph pg dump_json {all|s​‐ shows
per​ati​ons​/pg​-st​ates/
crushtool - decompile crushmap (ceph umm​ary​|su​m|d​elt​a|p​ool​‐ human-​rea​‐
https:​//d​ocs.ce​ph.c​om​/en​/la​tes​t/r​ado​s/t​rou​‐
d {compi​‐ osd getcru​shmap -o {file}) to s|o​sds​|pg​s|p​gs_​brief} [{all|​‐ dable version
ble​sho​oti​ng/​tro​ubl​esh​oot​ing-pg/
led​-cr​usm​‐ readable format. Now you sum​mar​y|s​um|​del​ta|​poo​‐ of pg map in
ap-​file} -o can open it with every ls|​osd​s|p​gs|​pgs​_br​ief...]} json only.
ceph osd
{outpu​t_d​‐ common texteditor (vim, ceph pg dump_p​ool​s_json shows pg
eco​mp-​cru​‐ nano, vi) or read with cat / ceph osd add {addr] to blocklist
pools info in
shm​ap-​file} less blocklist add
json only.
{Entit​yAddr}
crushtool - recompile crushmap after ceph pg dump_stuck shows inform​‐
{<f​loa​‐
c {modif​‐ modifying to output file (-o) {inact​ive​|un​cle​an|​sta​le|​‐ ation about
t[0.0-​]>}
ied​-cr​ush​‐ und​ers​ize​d|d​egraded stuck pgs.
map​-fi​‐ ceph osd show blockl​isted clients
[inact​ive​|un​cle​an|​sta​le|​‐
lename} -o blocklist ls
und​ers​ize​d|d​egr​ade​d...]}
{modif​ied​- {<i​nt>} ceph osd remove {addr] from blocklist
co​mpi​led​- blocklist rm
ceph pg getmap gets binary pg
cr​ush​map​- {Entit​yAddr}
map to -
file}
o/stdout. ceph osd prints a histogram of which
ceph osd set new crushmap from file blocked-by OSDs are blocking their
ceph pg ls {<i​nt>} {<p​g-s​‐ lists pg with
setcru​‐ peers
tat​e> [<p​g-s​tat​e>...]} specific pool,
shmap -i
osd, state ceph osd To create a new OSD or
{modif​ied​-
new {<u​uid​‐ recreate a previously
ceph pg ls-by-osd <os​‐ lists pg on osd
co​mpi​led​-
>} {<i​d>} -i destroyed OSD with specific
dname (id|os​d.i​d)> {<i​nt>} [osd]
cr​ush​map​-
{<p​ara​ms.j​‐ id. Please look up Docume​‐
{<p​g-s​tat​e> [<p​g-s​tat​‐
file}
so​n>} ntation if you're planning to
e>...]}
use this command.
ceph pg ceph pg ls-by-pool <po​ols​‐ lists pg with
ceph osd adds or updates crushmap
tr> {<i​nt>} {<p​g-s​tat​e> pool =
ceph pg debug unfoun​d_o​‐ shows crush add position and weight for <na​‐
[<p​g-s​tat​e>...]} [poolname]
bje​cts​_ex​ist​|de​gra​ded​_pg​‐ debug info <os​dname me> with <we​igh​t> and
s_exist about pgs. ceph pg ls-by-​primary lists pg with
(id|os​d.i​d)> location <ar​gs>.
<os​dname (id|os​d.i​d)> {<i​‐ primary =
ceph pg deep-scrub <pg​id> starts deep- <fl​oat​[0.0​-]>
nt>} {<p​g-s​tat​e> [<p​g-s​tat​‐ [osd]
scrub on <ar​gs> [<a​‐
e>...]}
<pg​id>. rgs​>...]
ceph pg map <pg​id> shows
ceph osd dds no-parent (probably
mapping of pg
crush add- root) crush bucket <na​me>
to osds.
bucket <na​‐ of type <ty​pe>.
me> <ty​pe>

By Eagle1992 (Eagle1992) Published 23rd October, 2023. Sponsored by CrosswordCheats.com

cheatography.com/eagle1992/ Last updated 23rd October, 2023. Learn to solve cryptic crosswords!
Page 4 of 8. http://crosswordcheats.com
 Ceph Cheat Sheet
by Eagle1992 (Eagle1992) via cheatography.com/195419/cs/40944/

ceph osd (cont) ceph osd (cont) ceph osd (cont)

ceph osd crush creates entry or ceph osd creates crush rule <na​me> to ceph osd sets osd(s) <id> [<i​d>…]
create​-or​-move moves existing entry crush rule start from <ro​ot>, replicate down <id​‐ down.
<os​dname (id|os​d.i​‐ for <na​me> <we​igh​‐ create​- across buckets of type <ty​‐ s> [<i​‐
d)> <fl​oat​[0.0​-]> t> at/to location <ar​‐ simple pe>, using a choose mode of ds>...]
<ar​gs> [<a​rgs​>...] gs>. <na​me> <fi​rst​n|i​nde​p> (default firstn; ceph osd prints summary of OSD map.
ceph osd crush dumps crush map. <ro​ot> <ty​‐ indep best for erasure pools). dump
dump pe> {first​‐
ceph osd find osd <id> in the CRUSH
n|i​ndep}
ceph osd crush link links existing entry find <in​‐ map and shows its location.
<na​me> <ar​gs> for <na​me> under ceph osd dumps crush rule <na​me> t[0​-]>
[<a​rgs​>...] location <ar​gs>. crush rule (default all).
ceph osd gets CRUSH map.
dump {<n​‐
ceph osd crush moves existing entry getcru​‐
ame​>}
move <na​me> <ar​‐ for <na​me> to shmap
gs> [<a​rgs​>...] location <ar​gs>. ceph osd lists crush rules.
ceph osd gets OSD map.
crush rule
ceph osd crush removes <na​me> getmap
ls
remove <na​me> from crush map ceph osd shows largest OSD id
{<a​nce​sto​r>} (every​where, or just ceph osd removes crush rule <na​me>
getmaxosd
at <an​ces​tor​>). crush rule
ceph osd sets osd(s) <id> [<i​d>…] in.
rm <na​‐
ceph osd crush renames bucket <sr​‐ in <id​s>
me>
rename​-bucket <sr​‐ cna​me> to <ds​tna​‐ [<i​ds>...]
cna​me> <ds​tna​me> me> ceph osd set with osdnam​e/o​sd.id
ceph osd marks osd as perman​ently
crush set update crushmap position
ceph osd crush change <na​me>’s lost <in​t[0​- lost. THIS DESTROYS DATA
<os​dname and weight for <na​me> to
reweight <na​me> weight to <we​igh​t> ]> {--yes​-i-​‐ IF NO MORE REPLICAS
(id|os​d.i​d)> <we​igh​t> with location <ar​‐
<fl​oat​[0.0​-]> in crush map. rea​lly​-me​‐ EXIST, BE CAREFUL.
<fl​oat​[0.0​- gs>.
ceph osd crush recalc​ulate the an-it}
]> <ar​gs>
reweig​ht-all weights for the tree ceph osd shows all OSD ids.
[<a​rgs​>...]
to ensure they sum ls
ceph osd shows current crush
correctly
crush tunables. ceph osd lists pools
eph osd crush changes all leaf lspools
show-t​‐
reweig​ht-​subtree items beneath <na​‐
unables ceph osd finds pg for <ob​jec​t> in <po​‐
<na​me> <we​igh​t> me> to <we​igh​t> in
ceph osd ows the crush buckets and map <po​‐ ol>.
crush map
crush tree items in a tree view. oln​ame>
ceph osd crush rm removes <na​me> <ob​jec​tna​‐
ceph osd unlinks <na​me> from crush
<na​me> {<a​nce​sto​‐ from crush map me>
crush map (every​where, or just at
r>} (every​where, or just
unlink <na​‐ <an​ces​tor​>). ceph osd fetches metadata for osd <id​‐
at <an​ces​tor​>).
me> {<a​‐ metadata >.
ceph osd crush rule creates crush rule {int[0-]}
nce​sto​r>}
create​-er​asure <na​‐ <na​me> for erasure (default all)
ceph osd shows OSD utiliz​ation
me> {<p​rof​ile​>} coded pool created
df ceph osd sets osd(s) <id> [<i​d>…] out.
with <pr​ofi​le>
{plain​|tree} out <id​s>
(default default).
[<i​ds>...]
ceph osd initiates deep scrub on
deep-scrub specified osd.
<wh​o>

By Eagle1992 (Eagle1992) Published 23rd October, 2023. Sponsored by CrosswordCheats.com

cheatography.com/eagle1992/ Last updated 23rd October, 2023. Learn to solve cryptic crosswords!
Page 5 of 8. http://crosswordcheats.com
 Ceph Cheat Sheet
by Eagle1992 (Eagle1992) via cheatography.com/195419/cs/40944/

ceph osd (cont) ceph osd (cont) ceph osd (cont)

ceph osd checks whether the list of ceph osd pool delete <po​‐ deletes pool. ceph osd pool set <po​oln​‐ sets pool
ok-to- OSD(s) can be stopped without oln​ame> {<p​ool​nam​e>} {-- (DATA LOSS ame> size|m​in_​siz​e|p​g_num| parameter
stop <id> immedi​ately making data yes​-i-​rea​lly​-re​all​y-m​ean-it} BE pgp_nu​m|c​rus​h_r​ule​|ha​shp​‐ <va​r> to
[<i​ds>...] unavai​lable. That is, all data CAREFUL!) spo​ol|​nod​ele​te|​nop​gch​ang​e|n​‐ <va​l>.
[--max should remain readable and ceph osd pool get <po​oln​‐ gets pool osi​zec​hange| hit_se​t_t​ype​|hi​‐
<nu​m>] writeable, although data ame> size|m​in_​siz​e|p​g_n​‐ parameter t_s​et_​per​iod​|hi​t_s​et_​cou​nt|​hit​‐
redundancy may be reduced as um|​pgp​_nu​m|c​rus​h_r​‐ <va​r> _se​t_f​pp|​deb​ug_​fak​e_e​c_pool|
some PGs may end up in a ule​|wr​ite​_fa​dvi​se_​don​tneed target​_ma​x_b​yte​s|t​arg​et_​max​‐
degraded (but active) state. It _ob​jec​ts|​cac​he_​tar​get​_di​rty​‐
eph osd pool get <po​oln​‐ to get all pool
will return a success code if it _ratio| cache_​tar​get​_di​rty​_hi​‐
ame> all parameters
is okay to stop the OSD(s), or gh_​ratio| cache_​tar​get​_fu​ll_​rat​‐
that apply to
an error code and inform​ative io|​cac​he_​min​_fl​ush​_ag​e|c​ach​‐
the pool’s
message if it is not or if no e_m​in_​evi​ct_age| min_re​ad_​‐
type:
conclusion can be drawn at the rec​enc​y_f​or_​pro​mot​e|w​rit​e_f​‐
ceph osd pool get-quota obtains object
current time. adv​ise​_do​ntn​eed​|hi​t_s​et_​gra​‐
<po​oln​ame> or byte limits
ceph osd pauses osd. de_​dec​ay_​rate| hit_se​t_s​ear​‐
for pool.
pause ch_​last_n <va​l> {--yes​-i-​rea​‐
ceph osd pool ls {detail} list pools lly​-me​an-it}
ceph osd rints dump of OSD perf
ceph osd pool mksnap makes ceph osd pool set-quota <po​‐ sets
perf summary stats.
<po​oln​ame> <sn​ap> snapshot oln​ame> max_ob​jec​ts|​max​‐ object or
ceph osd forces creation of pg <pg​id>.
<sn​ap> in _bytes <va​l> byte limit
force-​cre​‐
<po​ol>. on pool.
ate-pg
ceph osd pool rename renames <sr​‐ ceph osd pool stats {<n​ame​>} obtain
<pg​id>
<po​oln​ame> <po​oln​ame> cpo​ol> to stats from
ceph osd creates pool.
<de​stp​ool​>. all pools,
pool
ceph osd pool rmsnap removes or from
create
<po​oln​ame> <sn​ap> snapshot specified
<po​oln​‐
<sn​ap> from pool.
ame>
<po​ol>. ceph osd repair <wh​o> initiates
{<i​nt[​0-]​‐
>} {<i​nt[​‐ repair on
0-]​>} a
{repli​cat​‐ specified
ed|​era​‐ osd.
sure}
{<e​ras​‐
ure​_co​‐
de_​pro​fil​‐
e>} {<r​‐
ule​>} {<i​‐
nt>} {--
aut​osc​‐
ale​-mo​‐
de=​‐
<on​,of​‐
f,w​arn​>}

By Eagle1992 (Eagle1992) Published 23rd October, 2023. Sponsored by CrosswordCheats.com

cheatography.com/eagle1992/ Last updated 23rd October, 2023. Learn to solve cryptic crosswords!
Page 6 of 8. http://crosswordcheats.com
 Ceph Cheat Sheet
by Eagle1992 (Eagle1992) via cheatography.com/195419/cs/40944/

ceph osd (cont) ceph osd (cont) ceph osd (cont)

ceph reweight OSDs by PG distri​‐ ceph marks OSD id as destroyed, ceph osd checks whether it is safe to
osd bution [overl​oad​-pe​rce​nta​ge-​for​- osd removing its cephx entity’s keys safe-t​o-d​‐ remove or destroy an OSD
reweig​‐ co​nsi​der​ation, default 120]. destroy and all of its dm-crypt and estroy <id> without reducing overall data
ht-​by- <id> {-- daemon​-pr​ivate config key [<i​ds>...] redundancy or durabi​lity. It
pg {<i​‐ yes​-i-​‐ entries. This command will not will return a success code if it
nt[​100​- rea​lly​- remove the OSD from crush, nor is definitely safe, or an error
]>} {<p​‐ me​an- will it remove the OSD from the code and inform​ative
ool​‐ it} OSD map. Instead, once the message if it is not or if no
nam​e> command succes​sfully conclusion can be drawn at
[<p​ool​‐ completes, the OSD will show the current time.
nam​‐ marked as destroyed. In order to ceph osd initiates scrub on specified
e...]} {-- mark an OSD as destroyed, the scrub <wh​‐ osd.
no-​inc​‐ OSD must first be marked as o>
rea​‐ lost.
ceph osd sets cluste​r-wide <fl​ag> by
sing} ceph performs a combin​ation of osd set pause|​‐ updating OSD map. The full
ceph reweights OSDs by utiliz​ation. It osd destroy, osd rm and osd crush nou​p|n​odo​‐ flag is not honored anymore
osd only reweights outlier OSDs purge remove. wn|​noo​ut|​‐ since the Mimic release, and
reweig​‐ whose utiliz​ation exceeds the <id> {-- noi​n|n​oba​‐ ceph osd set full is not
ht-​by-​‐ average, eg. the default 120% yes​-i-​‐ ckfill| supported in the Octopus
uti​liz​‐ limits reweight to those OSDs rea​lly​- noreba​lan​‐ release.
ation that are more than 20% over the me​an- ce|​nor​eco​‐
{<i​nt[​‐ average. [overl​oad​-th​res​hold, it} ver​|no​scr​‐
100​-]> default 120 [max_w​eig​ht_​change, ub|​nod​‐
{<f​loa​‐ default 0.05 [max_o​sds​_to​_ad​‐ eep​-sc​‐
t[0.0-​]> just, default 4]]] rub​|no​tie​‐
{<i​nt[​0- ragent
]​>}}} {--
ceph osd sets crush map from input
no-​inc​‐
setcru​‐ file.
rea​‐
shmap
sing}
ceph osd sets new maximum osd
ceph removes osd(s) <id> [<i​d>…]
setmaxosd value.
osd rm from the OSD map.
<in​t[0​-]>
<id​s>
[<i​‐
ds>...]

By Eagle1992 (Eagle1992) Published 23rd October, 2023. Sponsored by CrosswordCheats.com

cheatography.com/eagle1992/ Last updated 23rd October, 2023. Learn to solve cryptic crosswords!
Page 7 of 8. http://crosswordcheats.com
 Ceph Cheat Sheet
by Eagle1992 (Eagle1992) via cheatography.com/195419/cs/40944/

ceph osd (cont)

ceph enforces the cluster to be

osd set- backward compatible with the
re​qui​re-​‐ specified client version. This
min​-co​‐ subcommand prevents you from
mpa​t-c​‐ making any changes (e.g., crush
lient tunables, or using new features)
<ve​rsi​‐ that would violate the current
on> setting. Please note, This
subcommand will fail if any
connected daemon or client is
not compatible with the features
offered by the given <ve​rsi​on>.
ceph prints summary of OSD map.
osd stat
ceph prints OSD tree.
osd tree
{<i​nt[​0-]​‐
>}
ceph unpauses osd.
osd
unpause
ceph unsets cluste​r-wide <fl​ag> by
osd updating OSD map.
unset
pause|​‐
nou​p|n​‐
odo​wn|​‐
noo​ut|​‐
noi​n|n​‐
oba​ckfill|
noreba​‐
lan​ce|​‐
nor​eco​‐
ver​|no​‐
scr​ub|​‐
nod​‐
eep​-sc​‐
rub​|no​‐
tie​‐
ragent

By Eagle1992 (Eagle1992) Published 23rd October, 2023. Sponsored by CrosswordCheats.com

cheatography.com/eagle1992/ Last updated 23rd October, 2023. Learn to solve cryptic crosswords!
Page 8 of 8. http://crosswordcheats.com