| PRIMECLUSTER Global Disk Services Configuration and Administration Guide 4.1 (Solaris(TM) Operating System) |
|
Contents
|
| Appendix F Troubleshooting | > F.1 Resolving Problems |
For EMC Symmetrix related errors, in one of the following circumstances, take action as indicated for the relevant situation.
The method handling master-to-proxy copying can be checked in the CPTYPE field that is displayed with the sdxinfo command. When the CPTYPE field displays soft without displaying SRDF, instead of SRDF, the GDS soft copy function is performing the copy operation.
|
# sdxinfo -S -e long -o pv1 OBJ NAME CLASS GROUP DISK VOLUME JRM MODE STATUS COPY CURBLKS COPYBLKS DLY CPTYPE CPSOURCE ------ ------------ ------- ------- ------- ------- --- ---- -------- ---- -------- -------- --- ------ -------- slice * Class1 Group1 Disk1 pv1 * * COPY run 93952 10027008 0 soft * |
The following five possible causes can create a situation where SRDF is not used.
For details, see "Using EMC TimeFinder or EMC SRDF in a Proxy Configuration."
1) Check whether the possible cause is (Cause a).
1-1) Identify the physical disk names of the disks connected with the master and the proxy groups.
|
# sdxinfo -D -o mg1 OBJ NAME TYPE CLASS GROUP DEVNAM DEVBLKS DEVCONNECT STATUS ------ ------- ------ ------- ------- ---------- -------- ------------- ------- disk Disk1 mirror Class1 mg1 emcpower10 17596416 node1:node2 ENABLE disk Disk2 mirror Class1 mg1 emcpower11 17682084 node1:node2 ENABLE # sdxinfo -D -o pg1 OBJ NAME TYPE CLASS GROUP DEVNAM DEVBLKS DEVCONNECT STATUS ------ ------- ------ ------- ------- ---------- -------- ------------- ------- disk Disk3 mirror Class1 pg1 emcpower20 17596416 node1:node2 ENABLE |
In this example, the physical disks emcpower10 and emcpower11 have been connected with the master group mg1, and the physical disk emcpower20 has been connected with the proxy group pg1.
If a lower level group is, or multiple disks or lower level groups are, connected with the proxy group, the possible cause is (Cause a).
1-2) Identify the Symmetrix IDs and the Symmetrix device names of the disks connected with the master and the proxy groups using the sympd command provided by SYMCLI.
|
# sympd list Symmetrix ID: 000183600262
Device Name Directors Device
--------------------------- ------------ --------------------------------------
Cap
Physical Sym SA :P DA :IT Config Attribute Sts (MB)
--------------------------- ------------ --------------------------------------
...
/dev/rdsk/emcpower10c 000 03A:0 02B:C0 RDF1 Grp'd RW 8592
/dev/rdsk/emcpower11c 001 03A:0 02B:D0 Unprotected N/Grp'd RW 8632
/dev/rdsk/emcpower12c 002 03A:0 01B:D0 RDF2 Grp'd RW 8632
...
Symmetrix ID: 000282600920
Device Name Directors Device
--------------------------- ------------ --------------------------------------
Cap
Physical Sym SA :P DA :IT Config Attribute Sts (MB)
--------------------------- ------------ --------------------------------------
...
/dev/rdsk/emcpower20c 005 03A:0 02B:C0 RDF2 Grp'd RW 8592
/dev/rdsk/emcpower21c 006 03A:0 01B:D0 RDF1 Grp'd RW 8632
/dev/rdsk/emcpower22c 007 03A:0 01A:D0 RDF1 Grp'd RW 8632
... |
In this example, as to the physical disks emcpower10c and emcpower11c, the Symmetrix ID is 000183600262 and the Symmetrix device names are 000 and 001 respectively. As for the physical disk emcpower20c, the Symmetrix ID is 000282600920 and the Symmetrix device name is 005.
For details on the sympd command, see the Symmetrix Manager manual.
1-3) Check the source (R1) device and the target (R2) device as the SRDF pair whether they are connected with the master and proxy groups respectively using the symrdf command provided by SYMCLI.
The SymDev field and the RDev field of the symrdf list command output display the Symmetrix device names of the source (R1) device and the corresponding target (R2) device respectively.
|
# symrdf list Symmetrix ID: 000183600262
Local Device View
-------------------------------------------------------------------------------
STATUS M O D E S RDF S T A T E S
Sym RDF --------- ------------ R1 Inv R2 Inv -----------------------
Dev RDev Typ:G SA RA LNK Mode Dom ACp Tracks Tracks Dev RDev Pair
--- ---- ----- --------- ------------ ------ ------ --- ---- --------------
...
000 005 R1:1 RW RW NR SYN DIS OFF 0 0 RW RW Synchronized
002 006 R2:2 RW RW NR SYN DIS OFF 0 0 RW RW Split
... Symmetrix ID: 000282600920
Local Device View
-------------------------------------------------------------------------------
STATUS M O D E S RDF S T A T E S
Sym RDF --------- ------------ R1 Inv R2 Inv -----------------------
Dev RDev Typ:G SA RA LNK Mode Dom ACp Tracks Tracks Dev RDev Pair
--- ---- ----- --------- ------------ ------ ------ --- ---- --------------
...
005 000 R2:1 RW RW NR SYN DIS OFF 0 0 RW RW Synchronized
006 002 R1:2 RW RW NR SYN DIS OFF 0 0 RW RW Split
... |
In this example, the source (R1) device 000 in the Symmetrix frame 000183600262 and the target (R2) device 005 in the Symmetrix frame 000282600920 are the SRDF pair. The master and the proxy have been configured properly, and the possible cause is not (Cause e).
For details on the symrdf command, see the Symmetrix Manager manual.
1-4) If the possible cause is (Cause e), release the master-proxy relationship once, and re-configure properly.
For example, according to the symrdf command results in step 1.3), you will see that the source (R1) device 006 in the Symmetrix frame 000282600920 and the target (R2) device in the Symmetrix frame 000183600262 are also the SRDF pair. The sympd command results in step 1.2) are indicating that the physical disk names of the Symmetrix devices 006 and 002 are emcpower21 and emcpower12 respectively. The proxy configuration associated with SRDF can be created by registering emcpower21 and emcpower12 with a class, connecting them with respective groups, and relating the groups as the master and proxy.
In addition, the registration of the source (R1) device with a device group can be done after registering the disk with a class, but it must be carried out before joining the master and the proxy.
2) Check whether the possible cause is (Cause b).
|
# sdxinfo -GV -e long -o pv1 OBJ NAME CLASS DISKS BLKS FREEBLKS SPARE MASTER TYPE WIDTH ------ ------- ------- ------------------- -------- -------- ----- ------ ------ ----- group pg1 Class1 Disk3 17596416 7503872 0 * mirror * OBJ NAME TYPE CLASS GROUP DISK MASTER PROXY SKIP JRM MODE CMODE LOCK 1STBLK LASTBLK BLOCKS STATUS PSLICE SNUM ------ ------- ------ ------- ------- ------ ------ ----- ---- --- ---- ----- ---- -------- -------- -------- -------- ------ ---- volume * mirror Class1 pg1 * * * * * * * * 0 65535 65536 PRIVATE * * volume pv1 mirror Class1 pg1 * mv1 * off on ro * on 65536 10092543 10027008 STOP on 1 volume * mirror Class1 pg1 * * * * * * * * 10092544 17596415 7503872 FREE * * |
In this example, the MASTER field of the group pg1 to which the proxy volume pv1 belongs displays the asterisk (*) and the MASTER field of pv1 displays the master volume name mv1, and the possible cause is (Cause b).
If the possible cause is (Cause b), the action as follows should be implemented.
2-1) Break the relationship between the master and proxy volumes.
|
# sdxproxy Break -c Class1 -p pv1 |
2-2) Join a pair of groups instead of a pair of volumes as the master and the proxy. In this example, the master group name is mg1.
|
# sdxproxy Join -c Class1 -m mg1 -p pg1 -a mv1=pv1:on |
3) Check whether the possible cause is (Cause c).
|
# sdxinfo -V -e long -o mg1 OBJ NAME TYPE CLASS GROUP DISK MASTER PROXY SKIP JRM MODE CMODE LOCK 1STBLK LASTBLK BLOCKS STATUS PSLICE SNUM ------ ------- ------ ------- ------- ------ ------ ----- ---- --- ---- ----- ---- -------- -------- -------- -------- ------ ---- volume * mirror Class1 mg1 * * * * * * * * 0 65535 65536 PRIVATE * * volume mv1 mirror Class1 mg1 * * * off on rw rw off 65536 10092543 10027008 ACTIVE on 1 volume * mirror Class1 mg1 * * * * * * * * 10092544 17596415 7503872 FREE * * # sdxinfo -V -e long -o pg1 OBJ NAME TYPE CLASS GROUP DISK MASTER PROXY SKIP JRM MODE CMODE LOCK 1STBLK LASTBLK BLOCKS STATUS PSLICE SNUM ------ ------- ------ ------- ------- ------ ------ ----- ---- --- ---- ----- ---- -------- -------- -------- -------- ------ ---- volume * mirror Class1 pg1 * * * * * * * * 0 65535 65536 PRIVATE * * volume pv1 mirror Class1 pg1 * mv1 * off on ro * on 65536 10092543 10027008 STOP off 1 volume * mirror Class1 pg1 * * * * * * * * 10092544 17596415 7503872 FREE * * |
In this example, the attributes of the physical slices of the master volume mv1 and the proxy volume pv1 are on and off respectively, and the possible cause is (Cause c).
If the possible cause is (Cause c), the action as follows should be implemented.
3-1) Break the master-proxy relationship.
|
# sdxproxy Break -c Class1 -p pg1 |
3-2) Relate the master and the proxy again. With the -a option for the sdxproxy Join command, if the physical slice attribute is not specified, the physical slice attribute of a created proxy volume conforms to that of the master volume.
|
# sdxproxy Join -c Class1 -m mg1 -p pg1 -a mv1=pv1:on |
4) Whether the possible cause is (Cause d) or (Cause e) can be determined by checking the command action history referencing the GDS log file /etc/opt/FJSVsdx/msglog/sdxservd.log.
If the possible cause is (Cause d) or (Cause e), the master and the proxy should be disjoined once and then rejoined.
TimeFinder and SRDF may stop operating because the exclusive lock on a Symmetrix device is not released due to an event such as a node down. In this event, copying by TimeFinder or SRDF does not operate normally, and the proxy volume status may become INVALID.
The exclusive locks for Symmetrix devices are divided into two types. Whether respective exclusive locks are left not released can be checked using the symdev and symcfg commands provided by SYMCLI.
|
# symdev -lock list |
If an exclusive lock has remained, the lock number is displayed.
Release the exclusive lock on the Symmetrix device with the following method.
As to an exclusive lock indicated by the symdev -lock list command, use the symdev command to release it. Substitute the lock number displayed with the symdev -lock list command for number.
|
# symdev -lock number release |
As to an exclusive lock indicated by the symcfg -lock list command, use the symcfg command to release it. Substitute the lock number displayed with the symcfg -lock list command for number.
|
# symcfg -lockn number release |
For details on the symdev command and the symcfg command, see the Symmetrix Manager manual.
|
Contents
|