Activity
From 2011-12-10 to 2012-01-08
2012-01-08
-
11:33 PM Bug #1945: Import GRUB 0.9.7 patches from Gentoo
- The mismatch patch wasn't uploaded for some reason. I am trying again.
-
- I discovered three issues in Illumos GRUB 0.97 when I tried porting it to Gentoo Linux. Patches are attached.
One ... -
10:27 PM Bug #1970 (Closed): Boot fails from an rpool on Hitachi 4k sector drives
- The drives are two Hitachi Desktar 7K1000 D drives in a mirrored pool. They are novel in that they have a single plat...
-
07:10 PM Bug #1946: incorrect formatting when listing output of multiple pools with zpool iostat -v
- I did publish the webrev on the mailing list already and got already a LGTM but for the sake of completeness here's t...
2012-01-07
-
05:50 PM Bug #453: ZPOOL_CONFIG_ASHIFT not set correctly for drives with large sector size
- confusion: i did not touch any bug report properties. not sure why my previous post shows that i changed *Tags* and *...
-
- where did the following lead to?
http://lists.illumos.org/pipermail/developer/2011-January/001218.html
quoting ... -
09:10 AM Bug #1961 (New): investigate stopping uniquifying CTF information
- We uniquify the CTF information in the kernel against genunix (and ip) to save space. This was probably really impor...
-
12:50 AM Bug #1960: developer tools should support git
- There is a perfectly usable one available in the OI SFE repository, though, which is the one I'm using. To get it:
...
2012-01-06
-
- (I could have sworn this was already filed, but can't find it)
Several of us (Me, Joshua, Joyent, Delphix, others.... -
08:46 PM Bug #1959 (Resolved): Assertion failed: p->key == key, file ../port/locale/collate.c, line 199, function substsearch in el_{CY,GR}.UTF-8 locales
- Collation data for el_{CY,GR}.UTF-8 seems to be broken, triggering the following assertion (simple testcase):...
-
10:01 AM Bug #1956: /usr/bin/uname '-o' should return 'illumos'
- GNU uname capitalises 'Solaris' so I think it's correct to do the same to Illumos. It also returns 'FreeBSD' and 'Li...
-
12:18 AM Bug #1956: /usr/bin/uname '-o' should return 'illumos'
- i think the correct name is "illumos" not "Illumos"
2012-01-05
-
05:29 PM Bug #1957: Printing to a Generic Postscript Network printer.
- I have tried removing the printer from CUPS and re-adding, closing all applications between to allow settings updates...
-
- Printing anything to a Lexmark C543dn printer the formatting is now off (prints too high up the page)
This printer... -
- Add new uname '-o' option to usage.
-
- The Illumos provided uname should support the '-o' argument and return 'Illumos' so that Illumos can be detected with...
-
03:11 PM Bug #1955 (Resolved): sderr kstat should expand when a LUN is expanded
- PSARC 2006/373 Dynamic Lun Expansion introduced the ability for LUNs to dynamically expand, however, the sderr kstat ...
-
- Running findleaks on a zfs crash show leaks in metaslab_group_create and zfs_ereport_checksum:
> ::findleaks -d
C... -
- The new C code for zfs allow calls zfs_get_fsacl() followed by a call to
zfs_set_fsacl(). The former uses zfs_ioctl(... -
- While investigating '1951 leaking a vdev when removing an l2cache device' we
uncovered another leak:
> ::findleak... -
- When removing a cache device, we reload the cache device list, dropping
the now unused devices. From spa_load_l2cac... -
- Adding -B option to ztest which takes the name of a directory which should be
an alternate root directory containing... -
- If the machine crashes during a spa reguid operation after the new guid has been synced to all labels, but before the...
-
- Historically the only way to see the usage of individual top-level devices within a pool was to run 'zpool iostat -v'...
-
- Resolved in r13555 commit:69e1739e786e
-
01:27 AM Bug #1439: Panic when booting on IBM SystemX
- As discussed the oi_151a live CD installs and runs on my IBM x3550 M3 with UEFI (BIOS) rev 1.13 in the default oi_151...
2012-01-04
-
- That doesn't remove the desire to have it stop making the kind of foolish mistakes that even simple compiler warnings...
-
07:45 PM Bug #1926: libresolv evades compiler warnings
- libresolv is years-old, years-obsolete library, replaced by libresolv2. It stays only for applications built 10+ year...
-
- first zpool iostat -v mpool (for the expected output)...
-
06:52 PM Bug #1917: ::umem_verify seg faults on a dump with rampant corruption
- No, this is a copy of a Joyent bug, in order to integrate a changeset that already exists
From e9af913034578289b... -
- Any chance to have reproducible testcase - the problematic corefile?
-
- pstop should do the same thing as other p-commands, something like:
pstop: cannot examine PID: process is traced
-
01:24 PM Bug #1778: Assertion failed: rn->rn_nozpool == B_FALSE, file ../common/libzfs_import.c, line 1077, function zpool_open_func
- Ok, another update... After trying to import it half a dozen times, the import succeeded. (Same thing happened last ...
-
- I hit this assertion again (with a different disk)......
-
- I wanted to merge ZFS support into Gentoo's GRUB, so I took the sources from the Illumos-Gate repository and then pro...
-
- Looking at the file ./usr/src/uts/intel/os/fmsmb.c it occurred to me that x86gentopo_legacy can be set in /etc/system...
-
- Hit the same issue on a brand new IBM x3550 M3 but on a newer BIOS, e.g. UEFI 1.13 (build date 9/23/2011) when doing ...
2012-01-01
-
03:09 PM Bug #1942 (New): scsi/mpt: Disconnected command timeout for Target X
- i have 3 sun servers running OSOL svn_111b. all of them have a LSI 1068E controller with each one j4200 disk enclosur...
-
07:06 AM Bug #1941 (Resolved): timer intervals incorrectly rounded to clock resolution
- If one currently attempts to create (say) a 1,000 usec POSIX interval timer
on x86, the actual cyclic created will n...
2011-12-30
-
- Resolved in r13553:3bed7426c561
-
- Resolved in r13554:fae070bfb73a
2011-12-29
-
- There are three instances in the gate of "sequence" being misspelt.
2011-12-28
-
06:32 PM Bug #1927: nfs4 stale clientid loop (OI server, Linux client)
- Just for information, we applied several fixes from Trond to resolve the problem on client side, even if it does not ...
-
05:50 PM Feature #644: Add Toshiba AL10/12/* devices to mpxio sym list
- See attached latest, greatest scsi_vhci.conf.
I've also added some comments to help the poor souls who might need ... -
- Updated patch which uses sys/isa_info.h to determine endianness.
-
- Here is my initial patch.
-
- "zfs get all" would be more useful if it supported the -t argument to limit its output only to a certain datatype.
2011-12-27
-
- The issue is that ZIL blocks should always be contiguous blocks and thus the "fast gang" logic should not apply to th...
-
02:50 PM Bug #417: Stale OpenOwner entries are not reaped for active clients
- Juho Juopperi wrote:
> Will the fix mentioned in http://lists.illumos.org/pipermail/developer/2010-December/001162.h... -
-
- smb_convert_wildcards incorrectly converts '<' at the end of the pattern to '*', making the results from @dir *.@ on ...
-
- smb_match{,_ci}_private functions incorrectly think that size of character == size of byte when handling wildcards:
...
2011-12-26
-
- Resolved in r13552:375fcb299d27
2011-12-25
-
- Hi Dan and Stepan,
Perhaps this is the same issue (or related) as Bug #1333 [https://www.illumos.org/issues/1333]
...
2011-12-24
-
- The same is true of the DTrace tests (baddof, at least)
2011-12-23
-
11:33 PM Bug #1774: keyboard does not work in kmdb at times when cold boot with "-v -m verbose -dk" appended to GRUB
- Spotted the same on x201
-
09:57 PM Bug #1928 (New): zfs send/receive kills COMSTAR
- We're running COMSTAR on oi_151a with the iSCSI LUNs backed by ZFS volumes. Whenever we try to back up the server, do...
-
02:50 PM Bug #1927 (Resolved): nfs4 stale clientid loop (OI server, Linux client)
- A filesystem mounted over nfs4 from an OI 151a server on a Linux client becomes unresponsive after running a filebenc...
-
11:46 AM Bug #1442: apic_timer_reprogram request is too far in future (on vmware workstation)
- I have the same issue on vbox 4.1.8.
-
- It's not evading them, it's just not specifying adding $(CCVERBOSE) to cflags when it, apparently, should. I'd thoug...
-
- Via a mechanism I haven't investigated, libresolv is built without cc -v, and thus with compiler warnings disabled. ...
2011-12-22
-
- Resolved in r13550 commit:84f53d070f62
-
- Resolved in r13550 commit:84f53d070f62
-
- Resolved in r13550 commit:84f53d070f62
-
- Resolved in r13551 commit:022a137cd76d
-
07:08 PM Bug #1925 (Resolved): stack overflow from mac code
- Joyent observed a panic with the following stack trace:...
-
- Here's some additional analysis from Dave Pacheco:
It's not clear that the state is invalid when we crashed. Regar... -
-
04:08 PM Bug #1918: stack overflow from mac_promisc_dispatch()
- Also, can someone with permission move this out of "site" and into "illumos kernel" or whereever this bug really belo...
-
- Simply moving the mac_perim_enter_by_mh() doesn't help. The aforementioned DTrace script still makes the machine sta...
-
- FWIW, here are the diffs for the dls_accept_promisc() solution:...
-
- I think I've been able to reproduce the bug.
Here's the scenario:
1.) Enable ip_forwarding.
2.) Consider thi... -
- Dan made the key breakthrough: ds_promisc is 0, but it's still on the mi_promisc_list. Looking at the code paths wher...
-
- Okay, I know what's wrong, but I'm not sure how to get into this state:
- ipnet snooping is *supposed* to never s... -
- sharefs is used to provide the list of active CIFS and NFS shares to user-space programms like sharemgr. Currently, s...
-
- Currently, the t_pcb of the panicking thread is plowed by the setjmp() in panicsys that saves the registers at the ti...
2011-12-21
-
- ...
-
- Running ::kgrep/::ugrep for sizes less than 4 (i.e., 1 or 2) doesn't properly work: it returns many too many hits.
-
- Increasingly in the system, timestamps are being recorded as TSCs (e.g., t_hrtime) – but this leaves them unscaled. S...
-
- We saw a double fault with the following stack trace:...
-
- Running ::umem_verify on a dump with corruption was seen to induce a core dump from mdb:
{code}
libumem.so`umem... -
-
- Man, no. Separate bugs for unrelated things.
-
- Corresponding with Joyent bugs:...
-
- ...
-
- Currently ipsec creates persistent kstats which are based upon the netstack id.
Having persistent kstats here does... -
06:51 PM Bug #1911 (Resolved): panic in smb_tree_connect
- ...
-
06:46 PM Bug #1873: comstar iscsi target; kernel panic idm; mutex_enter: bad mutex
- Dan McDonald wrote:
> Change both %x's in that line to %p.
Done; Here's:
Crash dump: http://ilavsky.name/oi/vm... -
- Change both %x's in that line to %p.
-
- Now the "iscsit_sess" script worked. I had to crash it from kmdb, system was hung.
Crash: http://ilavsky.name/oi/... -
- Dan McDonald wrote:
> I'm a little disappointed about the second one not catching anything. I may rewhack it to dep... -
- My bad on the modification. It's supposed to be iscsi*t*_conn_t:...
-
- Modified dtrace script is not workig, I got an error:...
-
- The ways to obtain a dump from a hung system are numerous, but fragile. The most reliable is...
-
02:33 PM Bug #1914 (Resolved): libses cannot set more than one properties at once
- If more than one properties are specified only the last one is set, other ones are discarded.
-
- Fixed patch: http://cr.illumos.org/view/bm9forx6
-
07:59 AM Feature #1910: Implement a generic dev_err
- Thank you for the comments.
Rich: I'm not sure how you'd implement a buffer-less version with vcmn_err. Could you ... -
- I asked for the macro to avoid the bogus static buffer. I'd be just as fine with use of a vprintf, vcmn_err, or what...
-
12:34 AM Feature #1910: Implement a generic dev_err
- I would prefer to avoid the macro version altogether. There are lots of reasons to deplore macros, and apart from si...
-
- FYI:
richlowe and I have cleaned up the web application powering the /msg URLs and it's on github here: https://g... -
12:34 AM Bug #1906: pkcs11_tpm will hang indefinitely if localhost:30003 is reachable
- Wyllys Ingersoll wrote:
> Im almost certain it is NOT fixed in the latest trousers package, I doubt that they are ev...
2011-12-20
-
- At present, zfs diff displays spaces in octal, as \040. Given it's a printable character, it seems entirely reasonab...
-
- The fix is pretty simple:...
-
- A stack backtrace is not an evaluation. It is, however, sufficient to see the severity of the problem. Write in the...
-
- [ Nexenta issue 6628 ]
Here's the stack back-trace:... -
- Now, I'd like to get people's input on this. Currently, I've got it implemented in my virtio drivers in a header file...
-
- A lot of drivers implement custom cmn_err wrappers that prepend the messages with "device:instance #". Implement a ge...
-
- From http://lists.illumos.org/pipermail/developer/2011-March/001651.html
> My understanding of the proposal was th... -
08:04 PM Bug #1906: pkcs11_tpm will hang indefinitely if localhost:30003 is reachable
- Im almost certain it is NOT fixed in the latest trousers package, I doubt that they are even aware that it is a problem.
-
- ...
-
- pkcs11_tpm is not the problem. The root of the problem is the TSS library - libtspi.so.1.1.3 - which comes from the ...
-
- Initially I discovered this while testing VMWare ESXi 4.1u2 on latest Illumos bits. I found out that the write perfor...
-
- I've a new theory, and a new D script to match:...
-
- Thank you. I wasn't expecting that much output. Also, I printed the wrong field vs. what actually crashed.
Can y... -
- I see .. delta is ~150 lines out of ~17k.
So the scenario now:
Fresh OI boot, original idm module, flags set, d... -
- The double-sized probably isn't code. Try this if you don't believe me:
dis orig-idm | wc -l
dis new... -
- Thanks, that indeed helped.
Dump is available here: http://ilavsky.name/oi/vmdump.6
Kmem flags are set as requ... -
- I have a DTrace script to try after you reset the system to its default idm module. You'll need to run it BEFORE the...
-
- Martin Ilavsky wrote:
>
> Is there a way to setup a watchdog that would reboot this system when it gets hung?
Th... -
- ..right, if I can't even spawn a login shell, hardly I would be able to spawn reboot (something I realized later, but...
-
- ...
-
- Can be closed as ACPICA update was integrated?
-
- Can this be closed, or re-targeted to current version (2011n)?
-
- in schpc pci_getslot_t->slot_condition is a 2-bit field, however in schpc_test we switch for PCIMSG_SLOTCOND_NOHOTPLU...
-
- Resolved in r13548 commit:c7b36cdbb672 (with just the check to disable PAD if CMT is also disabled).
2011-12-19
-
- You can make any consumer of crypto hang indefinitely by doing this:...
-
- Corresponds with Joyent issues
STOR-27 ZFS should be able to dump to a RAID-Z pool
STOR-57 new ZFS checksum... -
- Rich,
I just pulled new code (made sure I had no local changes) and built two new BE's to test your patch based on... -
- This is the latest version of Igor K's patch.
-
- Nevermind. I see we haven't added support for OpenSSL 1.0 yet.
-
- libkmsagent makes use of ERR_remove_state which has been deprecated in OpenSSL 1.0 but it also defines OPENSSL_NO_DEP...
-
10:53 AM Bug #1901: beadm fails to create a cloned BE
- truss output would be useful
-
- Upgraded to latest kernel (as of Friday 16th) from dev-il and the blocking issues seem to have "disappeared".
than... -
05:52 AM Bug #1884: Empty "used" field for zfs *space commands
- After looking at the concerned code and a few snips from the manpage.. I believe this is the expected behavior
Here...
2011-12-18
-
- ...
-
06:52 PM Bug #1898 (Resolved): Kernel assertion in zfs_putpage() raised by rfs_write()
- NFSv2 server side write() implementation raises kernel assertion in zfs_putpage().
panic[cpu0]/thread=ffffff017063... -
10:15 AM Bug #1896 (New): netstat shows duplicate LISTEN line for telnet and sunrpc
- #grep Build /var/adm/messages |tail -1
Dec 18 10:51:47 xxxxx genunix: [ID 365267 kern.notice] ^MOpenIndiana Build o...
2011-12-17
-
- Resolved in r13547 commit:7d365d7822b1
-
- Dan McDonald wrote:
> If you happen to be able to login on the console, doing "reboot -d" would be really helpful ... -
- Resolved in r13546 commit:6d7e5def82fb
2011-12-16
-
- Martin Ilavsky wrote:
> I used "-FnClmprt" flags. Yes, obj32/ was used. I turned kmem flags on before doing actual t... -
- I used "-FnClmprt" flags. Yes, obj32/ was used. I turned kmem flags on before doing actual tests.
Well your gut fe... -
- For a full build, -D *adds* a DEBUG build to the mix. Non-DEBUG is always there with nightly. I'm assuming you copi...
-
- Thanks, that helped. To do non-debug build, NIGHTLY_OPTIONS have to be set without "D", correct? I was not using scri...
-
- To use it on your running system you'll need to compile it with non-debug. A "make all" in $SRC/uts/intel/idm will c...
-
- I want to help, but I would need some guidance, if possible. I'm new to OI/illumos, I'm just getting familiar with it...
-
- I may have a lead (and Rich's denial-of-service possibility may be a deeper architectural flaw).
Look through here... -
- I'm assuming the "You" in the above is not me.
I don't think that reduces a crasher to a leak, I think it changes ... -
- Rich --> that appears to be the trigger. At the very least, I'd like to reduce this from a crasher to a memory leak....
-
- My assumption was that two clients with the same name was the trigger. Do we have anyway to notice that there's two ...
-
- Someone's releasing twice or forgetting to hold a second time, for sure. Question is, how to prevent it?
-
- Rich, if you wouldn't mind integrating it over the weekend, that'd be appreciated. That'll be sooner than we can get ...
-
- Joshua, I'd actually (for other reasons), be very interested in which parts of the /dev walk suck the most (perhaps b...
-
- We hit this internally at Joyent simply by having a large number of zones. What follows is from the internal analysis...
-
- Analysis here: https://gist.github.com/1484969
Suggested fix, thus far, is to tell ps not to look for ttys in a z... -
- Resolved in r13545 commit:b35304501ec3
-
- ...
-
- Talking to Trisk about this, I plan to fix this via the first part of the three things mentioned (only), that is disa...
-
- Debugging #1627, we end up with a PG hierarchy which is incomplete in a very strange way. Namely, we're missing enti...
-
- While debugging #1627 I happened upon the fact that (it appears) we're the only OS to actually trust the ACPI _CSD (a...
2011-12-15
-
- We have received updated driver code for the latest Emulex FC adapters. This is needed also to support the latest ge...
-
- As Garrett said "behavior is correct".
-
- WONTFIX
-
- I've set the kmem_flags (reboot) and set the initiator names on both clients to the same name. System crashed twice (...
-
- As of _PSARC/2009/663_, there is a new flag for zfs receive: *-e*.
It should be documented in zfs(1M). -
- Integrated in:
changeset: 13544:4fd590021359
tag: tip
user: Garrett D'Amore <garrett@nexenta.co... -
- The intel_nb500 code does this...
2011-12-14
-
- Dan McDonald wrote:
> 32-bit is okay now that I think I have a handle on WTF is going on. Can you please do the kme... -
- 32-bit is okay now that I think I have a handle on WTF is going on. Can you please do the kmem_flags trick for one m...
-
- Dan McDonald wrote:
> This is weird. The offsets you have (e.g. idm_conn_even_locked+0x7f) don't correspond to inst... -
- Dan McDonald wrote:
> Oh my... I just realized something. Are you running 32-bit by any chance?
Yeah, it's 32-bi... -
- Oh my... I just realized something. Are you running 32-bit by any chance?
Yes... yes you are. My bad. I'll contin... -
- This is weird. The offsets you have (e.g. idm_conn_even_locked+0x7f) don't correspond to instructions in my oi_151a ...
-
- Is is possible to upload the actual system dumps (vmdump.X files)? I'll be back with more in a bit.
-
- Huh, that's bad. Hey Dan, obvious implications.
-
- I had one of those facepalm moments yesterday ..
Both HPUX nodes were reignited (restored) from the golden image. ... -
- Resolved in r13543 commit:33634a52a6c1
Note that this fixes compilation, and brings cpr to the same state as _gcc ... -
- ...
-
- Resolved in r13542 commit:0657a48834a9
-
- Resolved in r13541 commit:7320fae78e0e
2011-12-13
-
- Probably this is an old problem with EFI labels reading
-
- Maxphys is a global variable. Let the user decide what is the maximum transfer size their driver supports.
2011-12-12
-
- And, sadly, assuming that coreID:powerdomain is 1:1 may break on AMD Bulldozer (I'm not sure I understand what's bein...
-
- Pushed in
changeset: 13540:97a48b559488
tag: tip
user: Alexander Eremin <a.eremin@nexenta.com... -
-
- Currently, cw does not shadow-compile things that aren't built with -c on the command line which means that we miss a...
2011-12-11
-
10:34 PM Bug #1772: OI cannot import a data zpool created by S11X
- If there's any chance that can help: the same pool is imported fine into FreeBSD 9.0rc3.
-
- Option #2 would require ctfconvert supporting multiple compilation units (I picked a bad test case to see whether it ...
-
- The use of TLS in certain ON libraries (libc, libses, libsmp) causes relocations in the DWARF debug information which...
-
- They didn't, and actually _removing_ it is a lot harder than you're giving credit to.
Moving it to nexenta's illum... -
- We almost certainly don't need ksh93, wonder if smartos did already drop it.
2011-12-10
-
11:37 PM Bug #1878 (Closed): mountd crashes during unsuccessful NFS mount
- I've been trying to set up OI 151a to share a ZFS folder over NFS to be used as an ESXi datastore. During my attempt...
-
-
- Further research has discovered two bugs that may have necessitated this, one of which is long since fixed, the other...
-
- usr/src/Makefile.ast forces ksh93 to use DWARF because "We need this for C99/VLA support", experimentally, that's not...
-
- At present ctfconvert treats any DWARF error as an absence of type information. That's often not true, the lack of a...
Also available in: Atom