similar to: sparc server plus zfs

Hi! We''ve started to poke and prod at Lustre 1.6.4.1, and it seems to mostly work (we haven''t had it OOPS on us yet like the earlier 1.6-versions did). However, we had this weird incident where an active client (it was copying 4GB files and running ls at the time) got evicted by the MDS and all OST''s. After a while logs indicate that it did recover the connection

Hi, I would like to know what TCP/UDP ports should i keep open in my firewall policies on my MGS server such that I can have my MGS server fire-walled. Also if in a event of loss of MGT would it be possible to recreate the MGT without loosing data or bringing the filesystem down (i.e. by using cached information from MDT''s and OST''s) Thanks Anand

Hi, We have seen a huge performance drop in 1.6.3, due to the checksum being enabled by default. I looked at the algorithm being used, and it is actually a CRC32, which is a very strong algorithm for detecting all sorts of problems, such as single bit errors, swapped bytes, and missing bytes. I''ve been experimenting with using a simple XOR algorithm. I''ve been able to recover

Hi all! I've been (very) out of the loop for a while, up to the point that I had completely missed that the mailing list has moved. Oh well. However, Kjell Claesson poked at me regarding the 3phase stuff, so I found out that I had missed the recent events. In any case, it seems that the situation is unchanged regarding 3phase support. As some of you remember I made a suggestion for a

Hi all! I propose to replace the variables output.voltage.target.line and output.voltage.target.battery with the variable output.voltage.nominal. The reasons for doing it: * Unify naming, the term "nominal" is used both in NUT and by UPS makers for this. * Very few devices (none that I have seen) have a nominal output voltage that differs when on battery/mains. * A single

The attached patch is a first stab at adding 3phase support to upsstats/upsimage. I got a little carried away, so bare with me. This patch adds 3phase awareness to upsstats/upsimage and the template files, and in additional to the obvious 3phase variables: - upsstats: * Adds IF/ELSE-awareness, and since the parser logic is rather limited this means that we now have IF, ELSE IF etc since all

Hi all! The following files emits warnings when compiled with gcc 4.0: al175.c bcmxcp_ser.c belkinunv.c cyberpower.c everups.c powercom.c solis.c All warnings seem to be of this variety: everups.c:38: warning: pointer targets in passing argument 2 of 'ser_get_char' differ in signedness I suggest that those who fiddles with those drivers fixes the warnings and verifies that it works

Hi, just wanted to confirm that Fiskars UPS 9000 series (aka Powerware Plus 500 series) with optional UPS Information Unit works fine with upscode2 driver. Tested on Fiskars 9200 (20kVA/400V/3ph). Regards, -- Sebastian Zagrodzki Dzia? Sieciowy ICM, Uniwersytet Warszawski s.zagrodzki at net.icm.edu.pl (+48-22) 5520527, 8268009, fax. 8284195 http://www.net.icm.edu.pl/

2006/1/27, Peter Selinger <selinger@mathstat.dal.ca>: > > ... > battery.mfr.date, which is probably the more accurate interpretation. > > I think the naming should be changed to propose both: - battery.date.mfr: Battery manufacturing date - battery.date.change: Battery change date Any comments? Arnaud -------------- next part -------------- An HTML attachment was scrubbed...

fellows, I'm working on integrating the preliminary HAL support, and thought a bit about the default configure values below: --with-snmp build and install SNMP drivers (no) --with-usb build and install USB drivers (no) [ --with-hal build and install HAL support (no) ] I wonder if some auto detection magic (if the requirement is present enable the

Hi fellows, As the NUT-NG draft will take me more time than I currently have (I found myself too optimistic since I got a baby ;-), below is the beginning of an answer to the "Common Power Management with UPS support" problem... First, NUT is pure userspace code, for portability issue, and will remain. USB is through libusb, snmp through net-snmp and serial through standard read/write.

Can somebody please explain why values are 0-prepended in almost every driver? Since it's done almost everywhere I suspect it's due to some old policy. In theory this might have been a good idea given the assumption that applications are forced to parse the values before presenting them. However, since this isn't done even within NUT I don't see any real reason for doing

Hi all! I'm wondering if there's anyone doing any hacking on the snmp-ups driver at the moment. I've started looking at implementing 3phase-support, and this uncovered some rather unpleasant stuff in the snmp-ups driver that I really need to fix in order to get things sane. Those of you that's allergic to SNMP might want to stop reading now, this is rather icky ;)

Hi! Can someone with CVS write access please remove generated files from the CVS tree? The ones currently on my annoyance-list are: Makefile data/html/Makefile Also, the following files (+more, I only compile a subset of the drivers) should be added to .cvsignore in the corresponding directories (they're all generated): config.log config.status clients/Makefile clients/upsimage.cgi

Hi again! Same setup as before (Lustre 1.6.2 + 2.6.18 kernel). This time things suddenly started to be very slow (as in periodically stalling), and we found a bunch of llog_ LustreErrors on the MDS. Some time later stuff had automagically recovered and is back to normal speed. Any idea on the meaning/cause of these errors? What are the seriousness of "LustreError" errors in

some notes: - for this kind of discussion, please cc upsdev list (I've added it), - I've fwded the full thread to upsdev, - this subject is part of what I call NUT-NG (next gen.), aka nut 3.0... - I'll study in depth the thread this week end, and complete with my thoughts, - thanks to Stan to have put us all in touch Arnaud 2006/3/10, Stanislav Brabec <sbrabec@suse.cz>: >

What is the process for integrating a Lustre+SAMFS solution into an existing customer environment. The plan is to have CRS build the Lustre component, but Lustre and SAMFS will need to configured and integrated into the customer computing environment. I am very familiar with the SAMFS integration, but not Lustre integration. Do we have resources in PS to provide the integration? Is this

Hello, I have been reading http://wiki.lustre.org/images/1/1b/Hadoop_wp_v0.4.2.pdf for setting up Hadoop over lustre. Generally in hadoop setup, we have 1 Namenode and various number of datanodes. If I want to setup the same keeping Lustre as backend, in the document it is mentioned that: ".............Our experiments run on cluster with 8 nodes in total, one is mds/namenode, the rest are

Hello, I am trying to mount a lustre filesystem thru an ib2ip gateway. The MDS''s have infiniband connections. The client nodes are tcp/ip connections. I am able to route between the client nodes and the MDS''s. I have the following in /etc/fstab: abe-mds1 at o2ib0,abe-mds2 at o2ib0:/home/client /abehome lustre _netdev,flock 0 0 I get the following when trying

on our cluster that has been running lustre for about 1 month. I have 1 MDT/MGS and 1 OSS with 2 OST''s. Our cluster uses all Gige and has about 608 nodes 1854 cores. We have allot of jobs that die, and/or go into high IO wait, strace shows processes stuck in fstat(). The big problem is (i think) I would like some feedback on it that of these 608 nodes 209 of them have in dmesg