October 2010 - Core-students - mailman3.cels.anl.gov

Where's Max?
by Max trefonides 26 Oct '10

26 Oct '10

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I have a doctor apt this morning, I'll be working from home until then and then I'll be at the lab this afternoon. - -- Max -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (Darwin) iEYEARECAAYFAkzGxPIACgkQaPBbFuCp8ULF0wCfZzQJ2wMoaolQj1lHoTNjCgq1 a58AnA96vSeVfSAPlARfdJ41jRjIh1vO =UWdl -----END PGP SIGNATURE-----

1 0

Output from "cron" command
by Super-User 25 Oct '10

25 Oct '10

Your "cron" job on mercury /root/bin/backup-nightly.sh | mailx -r systems(a)mcs.anl.gov -s "mercury Nightly Backup Cron" backup-admin(a)mcs.anl.gov produced the following output: cannot destroy 'zpool1/nmpdr-homes@nightly': dataset is busy cannot create snapshot 'zpool1/nmpdr-homes@nightly': dataset already exists

1 0

where's ken
by Kenneth Raffenetti 25 Oct '10

25 Oct '10

I'm going to take off early and get the scrape on the side of my car looked at. I'll check mail and be online later tonight. Ken

1 0

Fwd: Globus online endpoint (GridFTP) setup
by Craig Stacey 25 Oct '10

25 Oct '10

Begin forwarded message: > Hi Craig, > I enclose the draft instructions I have prepared for setting up a Globus online endpoint. I wonder whether someone in your group can follow these instructions to setup an endpoint on the machine that we identify for endpoint testing and provide feedback. This can be a starting point for your group to get familiarized with endpoint setup and we can use the feedback to improve the instructions before pointing the endusers at it. Please let me know. > > Thanks, > Raj == Craig

1 0

Re: Linux kernel hole in RDS.
by John Valdes 25 Oct '10

25 Oct '10

The full disclosure post about this is at: http://seclists.org/fulldisclosure/2010/Oct/283 I've verified that the sploit in that post works on Ubuntu. It doesn't seem to work on RHEL 5 even after making trivial mods so that it actually tries to run. RHEL 5 has an older kernel than those that are reportedly effected, but RH has backported RDS, so in principle, it may be vulnerable. John On Tue, Oct 19, 2010 at 04:24:29PM -0500, Gene Rackow wrote: > It appears there is yet another kernel hole in the wild. This time > the code directly impacts the linux kernel. > There is a kernel fix for it at the source level. I presume that > all the vendors will be coming up with fixes in the next couple days. > > Reliable Datagram Sockets (RDS) provide in order, non-duplicating, > highly available, low overhead, reliable delivery of datagrams between > hundreds of thousands of non-connected endpoints. > > This was code introduced in 2007, so is probably in everything currently > running. If anyone wants details and/or exploit code to try > contact me. > > The code does not appear to compile on MacOS or Solaris. It's a Linux > issue, not general unix. > > --Gene > > > > > _______________________________________________ > ihw_users mailing list > ihw_users(a)lists.anl.gov > https://lists.anl.gov/mailman/listinfo/ihw_users

3 6

a short term fix on the latest libc issues
by Gene Rackow 25 Oct '10

25 Oct '10

While this doesn't actually "fix" the problem and is a bit of security through obscurity, it does appear to work with no ill effects on the system. Testing indicates this can be done on a live machie without reboot. While whacking cron stops the initial script kiddie from getting root, it really doesn't fix the problem. There are a few other ways it can be exploited as well. Not sure if you want to use 5 spaces or some other random string. On Fri, 22 Oct 2010, Jim Mellander wrote: } We actually hotpatched the loader after testing (warning superdangerous): } } sed 's/AUDIT/ /' -i /lib64/ld-2.5.so } sed 's/AUDIT/ /' -i /lib/ld-2.5.so } } thats 5 spaces. } } } On Fri, Oct 22, 2010 at 5:26 PM, Jim Mellander <jmellander(a)lbl.gov> wrote: } > cron is but one of the numerous ways this can be exploited - i } > wouldn't count on cron permissions to prevent this. } > } > Imagine if a user can create a world-writable file owned by root into } > any directory that root can write in to - there's hundreds of ways } > this capability can be leveraged into gaining root privileges. �Thats } > what this exploit mechanism provides. �And, by the way, don't count on } > just the documented exploit - lots of .so's can be leveraged to } > exploit this. } > } > We're thinking of doing a binary hotfix to the loader & changing } > LD_AUDIT to another string (security thru obscurity) [...] ------- End of Forwarded Message

3 2

Where's Max?
by Max trefonides 25 Oct '10

25 Oct '10

Running a bit late. In around 9:30 -- Max - mobile

1 0

Where's Hunter
by Hunter Matthews 24 Oct '10

24 Oct '10

I've twisted my ankle pretty good. Please put me down for a sick day (which I guess technically should be an injured day). I should be online, BUT I would appreciate a call if anything major hits. -- Hunter Matthews Unix Administrator Office: Bldg 221 Room B240 Argonne National Labs, MCS Key: F0F88438 / FFB5 34C0 B350 99A4 BB02 9779 A5DB 8B09 F0F8 8438 Never take candy from strangers. Especially on the internet.

1 0

Output from "cron" command
by Super-User 24 Oct '10

24 Oct '10

Your "cron" job on mercury /usr/lib/fs/nfs/nfsfind produced the following output: find: stat() error /exports/proj-fischer00/fischer/pc/00000001.JPG: Value too large for defined data type find: stat() error /exports/proj-fischer00/fischer/pc/00000002.JPG: Value too large for defined data type find: stat() error /exports/proj-fischer00/fischer/pc/My Music/00000002.JPG: Value too large for defined data type find: stat() error /exports/proj-fischer00/fischer/pc/My Music/00000001.JPG: Value too large for defined data type find: stat() error /exports/proj-fischer00/fischer/pc/My Pictures/RR/DMIR_6/DMIR222a.JPG: Value too large for defined data type find: stat() error /exports/proj-fischer00/fischer/pc/My Pictures/RR/DMIR_6/DMIR222b.JPG: Value too large for defined data type find: stat() error /exports/proj-fischer00/fischer/pc2/00000001.JPG: Value too large for defined data type find: stat() error /exports/proj-fischer00/fischer/pc2/nek5n/doc/movies: Value too large for defined data type find: stat() error /exports/proj-fischer00/fischer/pc2/My Music/00000002.JPG: Value too large for defined data type find: stat() error /exports/proj-fischer00/fischer/pc2/My Music/00000001.JPG: Value too large for defined data type find: stat() error /exports/proj-fischer00/fischer/pc2/My Pictures/RR/DMIR_6/DMIR222b.JPG: Value too large for defined data type find: stat() error /exports/proj-fischer00/fischer/pc2/My Pictures/RR/DMIR_6/DMIR222a.JPG: Value too large for defined data type find: stat() error /exports/proj-fischer00/fischer/pc2/00000002.JPG: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/examples/contribtest/datatypes/bc6/bcr0c1.err: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/examples/contribtest/datatypes/bc1/bcr0c1.err: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/examples/contribtest/datatypes/bc5/bcr0c1.err: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/examples/contribtest/datatypes/bc10/bcr0c1.err: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/examples/contribtest/datatypes/bc2/bcr0c1.err: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/examples/contribtest/datatypes/bc8/bcr0c1.err: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/examples/contribtest/datatypes/bc0/bcr0c1.err: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/examples/contribtest/datatypes/bc7/bcr0c1.err: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/examples/contribtest/datatypes/bc9/bcr0c1.err: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/examples/contribtest/datatypes/bc3/bcr0c1.err: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/examples/contribtest/datatypes/bc11/bcr0c1.err: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/examples/contribtest/datatypes/bc4/bcr0c1.err: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/examples/perftest/long.err: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/examples/perftest/lrtestf.c: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/mpid/ch_nexus/nexus-rm/quickthreads/md/null.README: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/mpid/ch_nexus/nexus-rm/data_conversion/stamp-h.in: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/mpid/ch_nexus/nexus-rm/ports0/stamp-h.in: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/mpid/ch_nexus/nexus-rm/ports0/p0_th_null_i.h: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/mpid/ch_nexus/nexus-rm/ports0/p0_th_null.h: Value too large for defined data type find: stat() error /exports/shared/nt-homes001/lusk/mpich/mpid/ch_nexus/nexus-4.0b4/config/CONFIG: Value too large for defined data type

1 0

[CELS Sysadmins] The Plan
by Tisha Stacey 22 Oct '10

22 Oct '10

We have a reservation at Two Brothers' Taphouse for 7:00. It's under the name "Stacey". Of course, you don't have to wait until then to go. Just hang out at the bar; you can probably find something to drink. - Tisha _______________________________________________ This is the CELS Sysadmins mailing list. It is sent to memebers of: core-admins(a)mcs.anl.gov, hpc-admins(a)mcs.anl.gov, systems(a)alcf.anl.gov, aig(a)alcf.anl.gov. Do not subscribe directly to this list. Thanks.

1 0