Re: [ceph-users] Can't remove /var/lib/ceph/osd/ceph-53 dir
Try umount /dev/ - This should unmount the directory. Post this host restart may be needed to reflect the changes. Best Regards, Ranjit From: ceph-users [mailto:ceph-users-boun...@lists.ceph.com] On Behalf Of William Josefsson Sent: Tuesday, July 12, 2016 4:15 PM To: ceph-users@lists.ceph.com Subject: [ceph-users] Can't remove /var/lib/ceph/osd/ceph-53 dir Hi Cephers, I got problem removing /var/lib/ceph/osd/ceph-53 dir which was used by OSD.53 that I have removed. The way that I remove the OSD: 1. ceph osd out 53 2. sudo service ceph stop osd.53 3. ceph osd crush remove osd.53 4. ceph auth del osd.53 5. ceph osd rm 53 6. sudo umount /var/lib/ceph/osd/ceph-53 and then I try to 'sudo rm -rf /var/lib/ceph/osd/ceph-53' but got a following response: sudo rm -rf /var/lib/ceph/osd/ceph-53 rm: cannot remove '/var/lib/ceph/osd/ceph-53': Device or resource busy I use 'fuser -m' to check and saw following results: sudo fuser -m /var/lib/ceph/osd/ceph-53 /var/lib/ceph/osd/ceph-53: 1rce 2rc 3rc 5rc 8rc 9rc 10rc11rc12rc13rc14rc15rc16rc17rc18rc19rc 20rc21rc22rc23rc24rc25rc26rc27rc28rc29rc 30rc31rc32rc33rc34rc35rc36rc37rc38rc39rc 40rc41rc42rc43rc44rc45rc46rc47rc48rc49rc 50rc51rc52rc53rc54rc55rc56rc57rc58rc59rc 60rc61rc62rc63rc64rc65rc66rc67rc68rc69rc 70rc71rc72rc73rc74rc75rc76rc77rc78rc79rc 80rc81rc82rc83rc84rc85rc86rc87rc88rc89rc 90rc91rc92rc93rc94rc95rc96rc97rc98rc99rc 100rc 101rc 102rc 103rc 104rc 105rc 106rc 107rc 108rc 109rc 110rc 111rc 112rc 113rc 114rc 115rc 116rc 117rc 118rc 119rc 120rc 121rc 122rc 123rc 124rc 125rc 126rc 127rc 128rc 129rc 130rc 131rc 132rc 133rc 134rc 135rc 136rc 137rc 138rc 139rc 140rc 141rc 142rc 143rc 144rc 145rc 146rc 147rc 148rc 149rc 150rc 151rc 152rc 153rc 154rc 155rc 156rc 157rc 158rc 159rc 160rc 161rc 162rc 163rc 164rc 165rc 166rc 167rc 168rc 169rc 170rc 171rc 172rc 173rc 174rc 175rc 176rc 177rc 178rc 179rc 180rc 181rc 182rc 183rc 184rc 185rc 186rc 187rc 188rc 189rc 190rc 191rc 192rc 193rc 194rc 195rc 196rc 197rc 198rc 199rc 200rc 201rc 202rc 203rc 204rc 205rc 206rc 207rc 208rc 209rc 210rc 211rc 212rc 213rc 214rc 215rc 216rc 217rc 218rc 219rc 220rc 221rc 222rc 223rc 224rc 225rc 226rc 227rc 228rc 229rc 230rc 231rc 232rc 233rc 234rc 235rc 236rc 237rc 238rc 239rc 240rc 241rc 242rc 243rc 244rc 245rc 246rc 247rc 248rc 249rc 250rc 251rc 252rc 253rc 254rc 255rc 256rc 257rc 258rc 259rc 260rc 261rc 262rc 263rc 264rc 265rc 266rc 267rc 268rc 269rc 270rc 271rc 272rc 273rc 274rc 275rc 276rc 277rc 278rc 279rc 280rc 281rc 282rc 283rc 284rc 285rc 286rc 287rc 288rc 289rc 290rc 291rc 292rc 293rc 294rc 295rc 296rc 297rc 298rc 299rc 300rc 301rc 302rc 303rc 304rc 305rc 306rc 307rc 308rc 309rc 310rc 311rc 312rc 313rc 314rc 315rc 316rc 317rc 318rc 319rc 320rc 321rc 322rc 323rc 324rc 325rc 326rc 327rc 328rc 329rc 330rc 331rc 332rc 333rc 334rc 335rc 336rc 337rc 338rc 339rc 340rc 341rc 342rc 343rc 344rc 345rc 346rc 347rc 348rc 349rc 350rc 351rc 352rc 353rc 354rc 355rc 356rc 357rc 358rc 359rc 360rc 361rc 362rc 363rc 364rc 365rc 366rc 367rc 368rc 369rc 370rc 371rc 372rc 373rc 374rc 375rc 376rc 377rc 378rc 379rc 380rc 381rc 382rc 383rc 384rc 385rc 386rc 387rc 388rc 389rc 390rc 391rc 392rc 393rc 394rc 395rc 396rc 397rc 398rc 402rc 403rc 404rc 406rc 407rc 408rc 409rc 412rc 413rc 414rc 415rc 417rc 418rc 419rc 422rc 423rc 424rc 426rc 427rc 428rc 429rc 431rc 432rc 433rc 434rc 437rc 438rc 439rc 441rc 442rc 443rc 444rc 446rc 447rc 448rc 449rc 451rc 452rc 453rc 454rc 456rc 457rc 458rc 459rc 460rc 461rc 462rc 463rc 464rc 466rc 467rc 468rc 469rc 472rc 473rc 474rc 476rc 477rc 478rc 479rc 481rc 482rc 483rc 484rc 486rc 487rc 488rc 489rc 491rc 492rc 493rc 494rc 496rc 497rc 498rc 499rc 500rc 501rc 502rc 503rc 504rc 506rc 507rc 508rc 509rc 511rc 512rc 513rc 514rc 516rc 517rc 518rc 519rc 521rc 522rc 523rc 524rc 526rc 527rc
[ceph-users] Data recovery stuck
Hi All, I am in process of adding new OSDs to Cluster however after adding second node Cluster recovery seems to be stopped. Its more than 3 days but Objects degraded % has not improved even by 1%. Will adding further OSDs help improve situation or is there any other way to improve recovery process? [ceph@MYOPTPDN01 ~]$ ceph -s cluster 9e3e9015-f626-4a44-83f7-0a939ef7ec02 health HEALTH_WARN 315 pgs backfill; 23 pgs backfill_toofull; 3 pgs backfilling; 53 pgs degraded; 2 pgs recovering; 232 pgs recovery_wait; 552 pgs stuck unclean; recovery 3622384/90976826 objects degraded (3.982%); 1 near full osd(s) monmap e4: 5 mons at {MYOPTPDN01=10.115.1.136:6789/0,MYOPTPDN02=10.115.1.137:6789/0,MYOPTPDN03=10.115.1.138:6789/0,MYOPTPDN04=10.115.1.139:6789/0,MYOPTPDN05=10.115.1.140:6789/0}, election epoch 6654, quorum 0,1,2,3,4 MYOPTPDN01,MYOPTPDN02,MYOPTPDN03,MYOPTPDN04,MYOPTPDN05 osdmap e198079: 171 osds: 171 up, 171 in pgmap v26428186: 5696 pgs, 4 pools, 105 TB data, 28526 kobjects 329 TB used, 136 TB / 466 TB avail 3622384/90976826 objects degraded (3.982%) 23 active+remapped+wait_backfill+backfill_toofull 120 active+recovery_wait+remapped 5144 active+clean 1 active+recovering+remapped 104 active+recovery_wait 45 active+degraded+remapped+wait_backfill 1 active+recovering 3 active+remapped+backfilling 247 active+remapped+wait_backfill 8 active+recovery_wait+degraded+remapped client io 62143 kB/s rd, 100 MB/s wr, 14427 op/s [ceph@MYOPTPDN01 ~]$ Best Regards, Ranjit ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
Re: [ceph-users] Ceph-deploy new OSD addition issue
This is another error I get while trying to activate disk - [ceph@MYOPTPDN16 ~]$ sudo ceph-disk activate /dev/sdl1 2016-06-29 11:25:17.436256 7f8ed85ef700 0 -- :/1032777 >> 10.115.1.156:6789/0 pipe(0x7f8ed4021610 sd=4 :0 s=1 pgs=0 cs=0 l=1 c=0x7f8ed40218a0).fault 2016-06-29 11:25:20.436362 7f8ed84ee700 0 -- :/1032777 >> 10.115.1.156:6789/0 pipe(0x7f8ec4000c00 sd=6 :0 s=1 pgs=0 cs=0 l=1 c=0x7f8ec4000e90).fault ^Z [2]+ Stopped sudo ceph-disk activate /dev/sdl1 Best Regards, Ranjit +91-9823240750 From: ceph-users [mailto:ceph-users-boun...@lists.ceph.com] On Behalf Of Pisal, Ranjit Dnyaneshwar Sent: Wednesday, June 29, 2016 10:59 AM To: ceph-users@lists.ceph.com Subject: [ceph-users] Ceph-deploy new OSD addition issue Hi, I am stuck at one point to new OSD Host to existing ceph cluster. I tried a multiple combinations for creating OSDs on new host but every time its failing while disk activation and no partition for OSD (/var/lib/ceph/osd/ceoh-xxx) is getting created instead (/var/lib/ceph/tmp/bhbjnk.mnt) temp partition is created. The host I have is combination of SSD and SAS disks. SSDs are parted to use for Journaling purpose. The sequence I tried to add the new host as follows - 1. Ceph-rpms installed on new Host 2. from INIT node - ceph-disk list for new host checked 3. Prepared disk - ceph-deploy --overwrite-conf osd create --fs-type xfs {OSD node}:{raw device}, - Result showed that Host is ready for OSD use; however it didn't reflect in OSD tree (Because crush was not updated (?) ) neither /var/lib/OSD.xx mount got created. 4. Although it showed Host ready for OSD use; before it threw a warning that disconnecting after 300 seconds as no data received from new Host 5.I tried to activate the disk manually - a. sudo ceph-disk activate /dev/sde1 - This command failed to execute with erroneous values ceph-disk: Cannot discover filesystem type: device /dev/sda: Line is truncated After this I also tried to install ceph-deploy and prepare new host using below commands and repeated above steps but it still failed at same point of disk activation. ceph-deploy install new Host ceph-deploy new newHost Attached logs for reference. Please assist with any known workaround/resolution. Thanks Ranjit ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
[ceph-users] Ceph-deploy new OSD addition issue
Hi, I am stuck at one point to new OSD Host to existing ceph cluster. I tried a multiple combinations for creating OSDs on new host but every time its failing while disk activation and no partition for OSD (/var/lib/ceph/osd/ceoh-xxx) is getting created instead (/var/lib/ceph/tmp/bhbjnk.mnt) temp partition is created. The host I have is combination of SSD and SAS disks. SSDs are parted to use for Journaling purpose. The sequence I tried to add the new host as follows - 1. Ceph-rpms installed on new Host 2. from INIT node - ceph-disk list for new host checked 3. Prepared disk - ceph-deploy --overwrite-conf osd create --fs-type xfs {OSD node}:{raw device}, - Result showed that Host is ready for OSD use; however it didn't reflect in OSD tree (Because crush was not updated (?) ) neither /var/lib/OSD.xx mount got created. 4. Although it showed Host ready for OSD use; before it threw a warning that disconnecting after 300 seconds as no data received from new Host 5.I tried to activate the disk manually - a. sudo ceph-disk activate /dev/sde1 - This command failed to execute with erroneous values ceph-disk: Cannot discover filesystem type: device /dev/sda: Line is truncated After this I also tried to install ceph-deploy and prepare new host using below commands and repeated above steps but it still failed at same point of disk activation. ceph-deploy install new Host ceph-deploy new newHost Attached logs for reference. Please assist with any known workaround/resolution. Thanks Ranjit ___ ceph-users mailing list ceph-users@lists.ceph.com http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com