Hartmut Streppel
2009-Feb-11 15:46 UTC
[zfs-discuss] failmode= continue prevents zpool processes from hanging and being unkillable?
> Dear ZFS experts,> somehow one of my zpools got corrupted. Symptom is that I cannot > import it any more. To me it is of lesser interest why that happened. > What is really challenging is the following. > > Any effort to import the zpool hangs and is unkillable. E.g. if I > issue a "zpool import test2-app" the process hangs and cannot be > killed. As this originally happened on a Sun Cluster node, this > finally panic''ed the node, the storage resource failover to the other > node, hung again, panic''ed this node as well ..... Not a good scenario > for a production system. > > I now have 2 questions: > 1. Could the hang and the unkillable process be avoided by setting the > zpool''s failmode to continue? It is currently set to "wait" which I > think is the default. > 2. Is it true that I can only set the failmode property while the > zpool is imported? As trying to set it hangs and is unkillable. > > > If I do a "zpool import" the following shows up: > > pool: test2-app > state: ONLINE > scrub: none requested > config: > > NAME STATE READ WRITE CKSUM > test2-app ONLINE 0 0 0 > c1t1d7 ONLINE 0 0 0 > > errors: No known data errors > If I truss "zpool import test2-app" the last lines are: > open("/dev/dsk/c1t1d7s0", O_RDONLY) = 6 > fstat(6, 0xFFBFB404) = 0 > modctl(MODSIZEOF_DEVID, 0x01D80038, 0xFFBFB48C, 0x0008CC24, > 0xFF21B480) = 0 > modctl(MODGETDEVID, 0x01D80038, 0x0000001C, 0x000E9F90, 0xFF21B480) = 0 > fstat(6, 0xFFBFB404) = 0 > modctl(MODSIZEOF_MINORNAME, 0x01D80038, 0x00006000, 0xFFBFB48C, > 0xFF21B480) = 0 > modctl(MODGETMINORNAME, 0x01D80038, 0x00006000, 0x00000002, > 0x000F1FC8) = 0 > close(6) = 0 > ioctl(3, ZFS_IOC_POOL_STATS, 0xFFBFA4F0) Err#2 ENOENT > ioctl(3, ZFS_IOC_POOL_TRYIMPORT, 0xFFBFA5C8) = 0 > > > Thanks > Hartmut >-- Sun Microsystems GmbH Hartmut Streppel Sonnenallee 1 Systems Practice D-85551 Kirchheim-Heimstetten Phone: +49 (0)89 46008 2563 Germany Mobile: +49 (0)172 8919711 http://www.sun.de FAX: +49 (0)89 46008 2572 mailto: hartmut.streppel at sun.com Amtsgericht M?nchen: HRB 161028 Sun Microsystems GmbH, Sonnenallee 1, D-85551 Kirchheim-Heimstetten Gesch?ftsf?hrer: Thomas Schr?der, Wolfgang Engels, Dr. Roland B?mer Vorsitzender des Aufsichtsrates: Martin H?ring -- Sun Microsystems GmbH Hartmut Streppel Sonnenallee 1 Systems Practice D-85551 Kirchheim-Heimstetten Phone: +49 (0)89 46008 2563 Germany Mobile: +49 (0)172 8919711 http://www.sun.de FAX: +49 (0)89 46008 2572 mailto: hartmut.streppel at sun.com Sun Microsystems GmbH, Sonnenallee 1, D-85551 Kirchheim-Heimstetten Amtsgericht M?nchen: HRB 161028 Gesch?ftsf?hrer: Thomas Schr?der, Wolfgang Engels, Dr. Roland B?mer Vorsitzender des Aufsichtsrates: Martin H?ring
Reasonably Related Threads
- [LLVMdev] llvm-gcc-4.2 RELEASE_26 bootstrap failure on Solaris/SPARC - unhandled REAL_TYPE during compilation of '__powitf2'
- [LLVMdev] llvm-gcc-4.2 RELEASE_26 bootstrap failure on Solaris/SPARC - unhandled REAL_TYPE during compilation of '__powitf2'
- [LLVMdev] How to check for "SPARC code generation" in MachineBasicBlock.cpp?
- sw_trap_number from trap probe?
- [LLVMdev] llvm-gcc-4.2 RELEASE_26 bootstrap failure on Solaris/SPARC - unhandled REAL_TYPE during compilation of '__powitf2'