Activity
From 2013-06-30 to 2013-07-29
2013-07-29
-
09:01 PM Bug #3944 (In Progress): libtopo zfs module will clobber libzfs handle when an instance of the module unloads
-
- The problem here is that g_zfs global gets clobbered by other instances of the module doing a zfs_fini() - setting g_...
-
08:10 PM Bug #1101: NFS service fails to start after hostname change
- sham pavman wrote:
> Any opinions?
It should be fixed.
-
08:04 PM Bug #3943 (Closed): _snd_lowat_fraction tcp tunable has no effect
- From Sebastien Roy's original bug report at Delphix:...
-
- From Sebastien Roy's original bug report at Delphix:...
-
- ...
-
- ...
-
- ...
-
- ...
-
06:49 PM Bug #3934 (Resolved): project reference count leak in tasksys_settaskid()
- Resolved in 74bf729631d9843cabb29019f16ac648de4aaa80.
-
04:22 PM Feature #3778 (Resolved): 1000BASE-LX support in ixgbe
- Commit: eb341807e07639369211020c815f2b9002798a8a
https://github.com/illumos/illumos-gate/commit/eb341807e076... -
03:08 PM Bug #3845 (In Progress): beadm doesn't work in non-global zones
- webrev's updated, now zone root datasets from non-active global BE are marked with 'x' (if they are active on boot, t...
-
- Resolved in 69c7619071dc71a079b65cd199d237df0f487d15.
2013-07-28
-
- OpenSSH added a -m option to be able to extract a PEM formatted public key from an SSH public key. This is rather use...
-
- Specifying the -m option to prtconf allows both global and non-global zones to just get the amount of physical memory...
-
- So we're looking at some kind of crazy back up in SMF land:...
-
-
- Resolved in ca48f36f20f6098ceb19d5b084b6b3d4b8eca9fa.
-
- Resolved in 91948b51b8e978ddc88a36b2bc3ae83c20cdc9aa.
-
- This is duplicate for #1601
-
- From the Joyent bug report:
In Linux date(1) supports %N which outputs the nanoseconds of the current time:
<pr... -
03:27 PM Bug #3706 (Resolved): mdb segfault when random address is passed to "list" walker
- ...
-
-
- ...
-
-
- See 3936 for the analysis; opening this under a separate ticket because it is in fact a separate issue.
-
- We've seen several dumps in which the owner of cpu_lock is waiting for the cyp_rpend to go 0:...
-
- Currently, timer_settime() on a CLOCK_HIGHRES timer will remove any old cyclic associated with the timer, and create ...
-
- We have had quite a few dumps dying on seemingly random data corruption in the VM subsystem – always due to a mysteri...
-
- We saw several panics of the following form:...
-
- If vrrpadm's config file does not currently exist it will fail. Instead we opt to have it create the empty config fil...
-
- This is an enhancement such that vrrpadm can work with virtual nics which are commonly used inside of zones.
-
- The following is from the Joyent bug report....
-
- /usr/share/man/man1/tail.1 doesn't mentioned the '-F' option
-
- /usr/bin/tail -f FILE doesn't notice file truncation. /opt/local/bin/tail from pkgsrc is fine with it. Mac's tail is ...
-
- From the Joyent bug report:
yes(1) currently ignores all return values, which with a somewhat contrived example ma... -
- Take the following program:...
-
- From the Joyent bug report and evaluation:
The IP DCE is a cache of dce_t structures (see usr/src/uts/common/ine... -
- This privilege was added by Jerry back at Sun but it was done after the last manual page drop.
-
- Per Brendan Gregg:
root should be able to lower the nice value of a process within a zone.
renice -n −19 — shou... -
- Currently, libkvm (and mdb) require both a kernel crash dump (e.g., vmcore.x) and a kernel symbol table (e.g., unix.x...
-
- In thoth, some mdb processes were found to have died in v8.so:...
-
- libproc does some very odd things with respect to locating on-disk binaries that are represented in a core file, but ...
-
- ...
-
- Running "pargs -l" on an incomplete dump can result in the death of pargs itself:...
-
- We died in turnstile_block(), when we discovered that the mutex we were blocking on was, in fact, unheld:...
-
- From the Joyent bug report:
We died in bcopy() called from elfcore() (called from do_core()):... -
- From Keith's original bug description.
There are a small number of components on which illumos-joyent encodes ...
2013-07-27
-
07:39 PM Bug #3914: ill_frag_hash_tbl not allocated for loopback interfaces
- Digging in to the source... the call itself:...
-
- ...
-
- When you run *poweroff(1M)* on a */dev/term/** device, you get this quaint (and errant) message:...
2013-07-26
-
- A while back, gcc changed its expectations for the i386 ABI. Particularly, it changed it such that it is expected for...
-
- *Root Cause*
The problem is in the mdb_amd64_kvm_stack_iter() where there is the explicit check for recursion:
... -
- It looks like AMD64 is affected only by this bug.
-
- mdb truncates the displayed stack in a case there is a recursion detected. This is true for all stack displaying comm...
-
- *Root Cause*
The problem is that there is no check in the list walker for valid and consistent list_size and list_... -
-
03:01 PM Bug #3776: gcc can cause segfaults on ctfmerge
- The patch has been committed with minor changes in FreeBSD:
http://svnweb.freebsd.org/changeset/base/253661
2013-07-25
-
- T_ERROR is undocumented, ancient, and buggy.
T_ERROR should be completely removed from TLI. -
- Similar segfault could be caused by an uninitialized list:...
-
10:17 PM Bug #1986 (Resolved): libima misuses gethostname
- Resolved in 5d539a8
-
07:38 PM Bug #3909 (Closed): "zfs send -D" does not work
- As reported on the zfs@illumos mailing list:
The send completes (if piped to receive, the receive finishes OK), bu... -
- I'm hijacking this bug, because the fix is to have ZFS be more robust in the face of DKIOCGMEDIAINFOEXT failures.
2013-07-24
-
- It is not possible to access zfs snapshot via NFSv2:...
-
- It appears that dlstat(1M) is yet another command where we didn't get a manual page.
-
- The dlstat usage message states that -a/-A and -i may be used together, they may not.
It also fails to state that ... -
11:36 AM Bug #3594: Accessing a password protected CIFS share from an android device always fail
- Anyone is willing to help on investigating that one?
2013-07-23
2013-07-22
-
- Garrett D'Amore wrote:
> It would generally be good to know if this problem is specific to NFS client or server func... -
08:55 PM Bug #3858: Can't build nightly over NFSv4
- It would generally be good to know if this problem is specific to NFS client or server functionality.
-
05:48 PM Feature #3903 (New): DTrace SCTP Provider
- https://wikis.oracle.com/display/DTrace/sctp+provider
-
- Integrated:
commit 7256a34efe9df75b638b9e812912ef7c5c68e208
Author: Dan McDonald <danmcd@nexenta.com>
Date: ... -
- *Issue 4*
In a case the cv_wait() (or cv_timedwait()) will return prematurely in ipmi_submit_driver_request() we wil... -
- webrev is updated
-
02:56 AM Bug #3368: NFSv4 and slow OS X Mountain Lion (10.8.2) Client
- Hello,
I'd also like to ask if anybody has new information about this Apple problem... I know that is not caused b...
2013-07-21
-
- There are three issues related to ipmi_attach().
*Issue 1*
There is a race window between lines 128 and 129 her...
2013-07-20
-
10:48 PM Bug #3901 (New): incorrect behaviour of zfs receive -F
- _zfs receive -F_ does not respect the destinations canmount properties if it has to destroy/rollback some destination...
-
09:33 PM Bug #3900 (Resolved): illumos will not build against gcc compiled perl
- Previous title: cw doesn't understand gnu c options
The problem is that cw expects to get Studio options, includin... -
- A note by Irek Szczesniak:...
2013-07-19
-
08:01 PM Bug #3899 (New): VERIFY failure at space_map.c:519
- We observed this failure:
assertion failed: sm->sm_space == total (0x382fa90200 == 0x382fa8f600), file: ..
/../co... -
01:52 PM Feature #3897: zfs filesystem and snapshot limits
- Started by Bill and I during the fall 2012 hackathon.
-
- Provide a mechanism to limit how many filesystems or snapshots can be created beneath any point in the filesystem hie...
-
https://github.com/illumos/illumos-gate/commit/3e7c6556c4272555ed6c5ecdbca2bf8a56df6ff5-
-
- https://github.com/illumos/illumos-gate/commit/452bd827089206a0c637b3944aa91806f17304d7
-
- Will have to replumb the ndd setting functions for the aforementioned variables.
-
01:26 AM Bug #3891 (Closed): Incorrect Makefiles for C++ SPARC
- I found the cause of this linking failure. The symbolic links to C++ shared objects in /usr/lib were all broken in /...
2013-07-18
-
- The anonymous port ranges in TCP and UDP should be checked either at NDD setting time, or reset to sane during calls ...
2013-07-17
-
- The obvious solution here is simply to prohibit taking a snapshot of an inconsistent dataset and return EBUSY in this...
-
- Upon receiving a seemingly complete zfs snapshot from its leader, the mounted snapshot shows "phantom" file entries.
... -
06:59 PM Bug #3893: lvm: incorrect flag handling
- @
From 2c149996cafb19ac2a9fd81ec7f3e951037413c4 Mon Sep 17 00:00:00 2001
From: Prasad Joshi <pjoshi@stec-inc.com>
... -
- Copying the patch again
@
diff -u -p a/usr/src/uts/common/io/lvm/mirror/mirror.c b/usr/src/uts/common/io/lvm/mirr... -
- In lvm mirror write strategy function (mirror_write_strategy) a condition which checks for resync generated i/o is i...
-
- Resolved in 5c069a6
2013-07-16
-
- True, "zfs upgrade" is a workaround. -An even better workaround would be to unmount & remount the dataset, which is ...
-
- zfs upgrade old/fs will solve this. Is this even a bug in zfs(1M), or is it a bug in the python thing that it's auto...
-
- Pools upgraded from version 14 or earlier should be able to use user/group accounting (e.g. "zfs userspace"), however...
-
- Static linking! That's something I never considered. I see from `ldd' that it's clearly the problem. Even 64-bit C...
-
- I'm a bit surprised by that, since my builds on SPARC succeed. The order of dynamic libraries relative to objects on...
-
- In building illumos-gate on OpenSXCE2013.05 SPARC, I ran into problems with C++ compiles. This error only affects 32...
-
- Matt (our intern) has it under test now. We used this DTrace one-liner:
dtrace -n 'tcp_update_next_port:return / ar... -
11:22 AM Feature #3890 (New): New deduplication hash algorithm
- Hello guys,
do you think it's possible to add a new deduplication hash algoritm less cpu intensive?
I mean sha 25...
2013-07-15
-
- This happens when we disassemble 'enc_tab', which is data rather than text (though is in .text).
It is likely that... -
- So continuing on we still have the more important question of where did zfs go awry. If we look again at our header b...
-
- As best as I can determine, the mapfile v2 language has no strict equivalent to the old "?E" flag, to state that a se...
-
- When 'zfs recv -F' is used with an incremental recv it rolls back any changes made since the last snapshot in case ne...
-
05:17 PM Feature #3842: remove the machid commands
- If you want to remove the machid commands you'll also need to update /etc/profile since it uses /bin/i386 to determin...
-
- While bringing up a certain blkdev driver, I encountered problems with this as well (mostly pertaining to bad EXPANDS...
-
- Filer is correct in diagnosing the problem.
We have a summer intern here for a while, this is a perfect first Illu... -
- commit dba604f95b3742920553aafd90c5b67721b07723
Author: Dan McDonald <danmcd@nexenta.com>
Date: Wed Jul 10 10:58:... -
- Enlarge the file buffer to 64k in digest/mac for better performance.
Similar to this commit:f2728256f3ab94f4fc6667...
2013-07-13
-
- Most other platforms support a number of different codes in their /etc/issue file to print of things like hostname an...
-
- Attaching patch which uses libbe to mount and unmount boot environments. It passes cstyle and seems to work okay, bu...
-
- halt execs beadm to mount and unmount boot environments which is stupid when libbe provides a simple interface for do...
-
- Resolved in 01f9868a
Also, we don't need the kernel command line of _this_ boot, because:
1) it might be anothe...
2013-07-12
-
- We have another dump that seems highly related. Here we are in the midst of doing a rollback and a thread in question...
-
03:08 PM Bug #3882 (Resolved): Remove xmod & friends
- Once upon a time, US export restrictions forced Sun to dice up the source code for export. Since then, the restricti...
-
- I've also discovered another problem. After activating the new boot environment the zone cannot be started. It give...
-
- I applied the following patch in oi-hipster. It seems to fix complex.h (at least imagemagick can use it).
2013-07-11
-
- Andrew Stormont wrote:
> I've attached your patch with a couple of changes. It doesn't unset the boot environment i... -
- I've attached your patch with a couple of changes. It doesn't unset the boot environment if it's already set, and it...
-
- There is a bug. If you run activate on the same boot environment twice it will set the org.opensolaris.libbe:active ...
-
- Andrew Stormont wrote:
> OK, so the problem was with something else in my environment (I was running a new libzfs on... -
- OK, so the problem was with something else in my environment (I was running a new libzfs on an old kernel). After fi...
-
- I think this is yet another thing that is, right now, in libm (and thus not illumos)
I _think_ I actually fixed th... -
- The following test case works with Sun Studio cc but fails with gcc (checked illumos-gcc and 4.7):
$ cat test_comple... -
- It would be nice if there were a device driver for devices like the P410 and P420, since it's currently impossible to...
-
-
- Integrated to illumos-gate:...
-
- Sebastian Feld wrote:
> Wouldn't it be a good idea to disable NFSv4 support by default until it becomes stable enoug... -
11:51 AM Bug #3858: Can't build nightly over NFSv4
- Bug #3879 has been filed to disable NFSv4 support until it becomes stable.
-
- It happens for NFSv4.
Wouldn't it be a good idea to disable NFSv4 support by default until it becomes stable enoug... -
- Feel free to run:...
-
- Per https://www.illumos.org/issues/3858 the NFSv4 support is unstable and doesn't even support building Illumos itsel...
2013-07-10
-
10:15 AM Feature #3870: Request driver support for Marvell 88SE94[48]5 SAS HBA
- Coy Hile wrote:
> It's possible that all that we need here is to add entries to /etc/driver_aliases for the new PCI ... -
- ZFS uses the aforementioned ioctl(), but blkdev, while supporting little brother DKIOCGMEDIAINFO, doesn't support the...
2013-07-09
-
- The patch causes beadm to segfault in my build zone, caused by a call to abort in make_dataset_handle_common. Not su...
-
- Provided fix (webrev is available at http://cr.illumos.org/~webrev/alhazred/3845) changes libbe to allow
view, creat... -
- I've seen this as well on oi_151a7 on a 24-way machine. By disabling 22/24 cores, I was able to not see this bug. I...
-
- It's possible that all that we need here is to add entries to /etc/driver_aliases for the new PCI IDs. I'm testing t...
-
- Attempting to force-unmount the stale filesystem doesn't work, so avoiding the dmu_objset_own() EBUSY is of little va...
2013-07-08
-
10:35 PM Bug #3877: fast reboot does not work with extended partition
- Parsing of GRUB1 menu looks like ultimate crap.
Long term solution is to keep kernel command line inside the kerne... -
-
- When the system is installed in an extended partition, the fast reboot does not work:...
-
-
- Unfortunately, the unmount path does in fact use ZFS_ENTER(), specifically in:...
-
- One possible way to reproduce this:
Part 1 (race exploder):... -
- A system panicked with the following stack:...
-
- ...
-
- ...
-
- ...
-
03:36 PM Bug #3853: __cplusplus change and headers incompatibility with clang
- Some history:
malloc.h:
malloc, free, realloc, and calloc were declared by <malloc.h> in POSIX-1988 and the related X... -
- ...
-
- ...
-
- ...
2013-07-07
-
- Hello guys i just create a ramdisk with the ramdiskadm command,
i create a zpool on it and i setup my zfs properti... -
- Ian Collins wrote:
> Attempting to build SmartOS over NFS, my failure (same on two runs) was here:
>
> (cd /mnt/s...
2013-07-06
-
08:10 PM Bug #3871: fix issues introduced by 3604
- Igor Pashev wrote:
> 2013. Time to use C99. C99 is good :-)
There are two problems here. One is that the use of a va... -
- 2013. Time to use C99. C99 is good :-)
-
- GCC 4.5.3 on Gentoo Linux did not like a few of the changes made in the issue 3604 patch. It printed an error and a c...
-
- I'd like to request that support be added for the Marvell 88SE9485 SAS controller (aka Supermicro AOC-SAS2LP-MV8). I...
-
09:31 AM Bug #3858: Can't build nightly over NFSv4
- Attempting to build SmartOS over NFS, my failure (same on two runs) was here:
(cd /mnt/smartos-live/proto.strap &&...
2013-07-04
-
- The ZFSOnLinux project changed this code to use SLAB allocation when it was imported. Linux has 8KB stacks, so stack ...
-
- Closed at submitters request.
-
03:41 PM Feature #3399 (Feedback): Provide Xen 4.2/4.3 kernel support
- Please, close this ticket. Can be reviewed by the dependent parties.
-
02:52 PM Bug #3869 (Resolved): tcp anon port calculation wrong
- If tcp_largest_anon_port gets tuned down from the default of 65536, some percentage of all source ports for outgoing ...
2013-07-03
-
- ...
-
- The original bug report at Delphix:...
-
- ...
-
- From Jeff Biseda's bug report at Delphix:...
-
- Analysis by Jeff Biseda of a panic we hit at Delphix:...
-
- ...
-
02:24 PM Bug #3865 (New): AHCI: 88SE9230 watchdog timeouts, "hung" panic
- I've been running a box with a Marvell 88SE9230 on a kernel including the recent patches for the 88SE9128. The two ch...
2013-07-02
-
- From Jeremy Jones's bug report at Delphix:...
-
- From Jeff Biseda's bug report at Delphix:...
-
- We were hitting this repeatably on a customer system. After extensive investigation by Jeremy Jones and Jeff Biseda, ...
-
10:25 PM Bug #3858: Can't build nightly over NFSv4
- My guess would be race-condition type of things between parallel make jobs - at least, I have seen similar behavior w...
-
12:00 PM Bug #3858: Can't build nightly over NFSv4
- I build ON on a NFSv3 mountpoint without any issues.
-
- From the original bug at Delphix....
-
-
- Igor, thanks for spotting that. Yuri was right. I neglected to notice that ZFSOnLinux had changed this to be a dynami...
-
- Yuri Pankov wrote:
> There's nothing wrong with using sizeof() that way here.
Using sizeof on an dynamic allocati... -
- Yuri Pankov wrote:
> There's nothing wrong with using sizeof() that way here.
Unless path is a pointer:
https:... -
10:05 AM Bug #3860: ztest_vdev_aux_add_remove should use MAXPATHLEN instead of sizeof in snprintf()
- There's nothing wrong with using sizeof() that way here.
-
- The title has a typo. It should be snprintf(), not sprintf().
-
- GCC 4.8.0's new diagnostics detected a programming error in ztest.c when building ZFSOnLinux with on Gentoo Linux. I ...
2013-07-01
-
- Josef Sipek wrote:
> My guess is that this is a build system issue not an NFS client/server issue.
IIRC, there we... -
- My guess is that this is a build system issue not an NFS client/server issue.
-
- Trying to build the gate over NFS fails with:...
-
- Yes. The 88SE9128 should work now. Closing.
-
05:23 PM Bug #993 (Feedback): Marvell SATA controller 88SE9128 - driver issue?
- Marcel - Is this ticket fully resolved with https://www.illumos.org/issues/3815 ??
-
- I'm uploading an updated workaround for the current gate revision (14068:2547a41b1162).
2013-06-30
-
11:04 PM Bug #3855 (Feedback): I/O to pool appears to be hung - kernel panic
- After additional troubleshooting, there was a bad hard disk that was undetected by the OS. I believe when that speci...
-
- Coredump relating to the @panic message: I/O to pool 'export' appears to be hung.@ message;...
-
- Coredump relating to the @panic message: BAD TRAP: type=e (#pf Page fault) rp=ffffff00b9be9970 addr=20 occurred in mo...
-
- Experiencing intermittent kernel panics, first couple of panic messages were showing;
@panic message: BAD TRAP: t... -
- Last time I've chosen old patches. Correcting...
-
- One real issue was to reimport all declarations back to global namespace.
-
- Linux (glibc) still provides malloc.h. FreeBSD malloc.h errors out if __STDC__ is defined. LLVM errors out if malloc....
-
- I'm not sure about "no one using" - at least during clang compilation it was used in some way. I can see it is used u...
-
- No one should be using malloc.h any more, maybe just remove it?
-
- It would be useful to have the version of SEEPROM configuration in the devinfo tree for igb and ixgbe
-
- Change was integrated.
-
- The SUNW_NO_UPDATE_NOTIFY=1 should be added to env files to disable Sun Studio update checks.
Also available in: Atom