By the way, I've noticed that under QuickTools management media for ISL ports are shown as unknown. Is this normal? I must note that ISL XPAK cables are original and came from Qlogic.]]>
[...] Unfortunately, this didn't help completely. I still have one volume restarting (from 3). This time it's different error: Nov 23 12:54:05 xserve3 fsm[50668]: fsm(50668,0xf58dd000) malloc: *** error for object 0x5: Non-aligned pointer being freed\n*** set a breakpoint in [...]]]>
That's the most exciting message on here in a while!]]>
http://imvf.com/01boy-search.html http://imvf.com/0hala0faisal-search.html http://imvf.com/1000pix-search.html http://imvf.com/1001movies-search.html http://imvf.com/100calcinhas-search.html http://imvf.com/100tosdefotos-search.html http://imvf.com/101lolita-search.html http://imvf. [...]]]>
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*- FASTER! FASTER! FASTER! http://tryfreeporn.info/vids/vid01.avi http://tryfreeporn.info/vids/vid02.avi http://tryfreeporn.info/vids/vid03.avi http://tryfreeporn.info/vids/vid04.avi http://tryfreeporn.info/vids/vid05. [...]]]>
test]]>
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*- FASTER! FASTER! FASTER! http://adultsexkey.com/vids/vid01.avi http://adultsexkey.com/vids/vid02.avi http://adultsexkey.com/vids/vid03.avi http://adultsexkey.com/vids/vid04.avi http://adultsexkey.com/vids/vid05.avi *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*- [...]]]>
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*- FASTER! FASTER! FASTER! http://adultsexkey.com/vids/vid01.avi http://adultsexkey.com/vids/vid02.avi http://adultsexkey.com/vids/vid03.avi http://adultsexkey.com/vids/vid04.avi http://adultsexkey.com/vids/vid05.avi *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*- [...]]]>
Hi, We eventually were able to locate the Meta data - yes, it had been placed on a seperate xserve RAID system that had been disconnected and moved off the rack. Once we located the meta data awe were able to get things up and running again no problem. Thanks for all the help! [...]]]>
[...] I demoted replica as standalone and then binded htem to OD master. But the strange followinf message is back now : Nov 20 11:40:32 alpha servermgrd[66]: -[AccountsRequestHandler(AccountsOpenDirectoryHelpers) openLocalLDAPNodeIfNeeded]: dsLocalLDAP = (null), error = Error Domain=com. [...]]]>
Hello Thanks for the replies. The migration went smooth, managed to changed the 8 ip's i needed (considering the v-tracks).
One thing i did that made everything smooth was setting up an extra dns server with the new ip's so that the xsan admin app would be able to authenticate on the 3 servers as they differ on ip's. This had to be done so that the failover could be achieved.
It was fun! cheers On Wed, Nov 18, 2009 at 2:33 PM, David S. Long <email@hidden> wrote:
Assuming; 1) They didn't setup an "Xsan Metadata Controller" when configuring their MDCs. 2) Still need to follow standard rules for changing IP addresses on Mac OS X server, ie changeip and DNS.
You should be fine with following the admin guide.
David S. Long
Founder
Install • Migrate • Manage
web: Â http://www.v2consulting.com/
office: (415) 989-9889
email:Â email@hidden
ichat:Â email@hidden
On Nov 18, 2009, at 9:21 AM, Raul Basurto Rosenzweig wrote: > quoted text
]]>Assuming; 1) They didn't setup an "Xsan Metadata Controller" when configuring their MDCs. 2) Still need to follow standard rules for changing IP addresses on Mac OS X server, ie changeip and DNS. You should be fine with following the admin guide. David S. [...]]]>
Raul, These were exact procedures when I had to move all Xsan clients to a different subnet. It went pretty smoothly. I also wanted to move to a different metadata subnet but it turned out the safest way to do that was to rebuild the SAN from scratch. [...]]]>
Hello all I was just called to change the ip settings of the public network of a small xsan solution for a client. The company is going through some restructuring and they need those ip's changed completely.... another subnet, different mask, all of it. [...]]]>
[...] longer supports management with SanBox. You can only manage it with the Quicktool web GUI or the Enterprise Fabric Suite. My 10G are on default settings. OK I have updated firmware to 7.4.X on both switches, reset them and reapplied all port and zoning settings according Apple docs. [...]]]>
You should check switch firmware and make sure device path, LUN numbers, or WWN doesn't change when using ISL. Also, you should try to go back to your original setting and see if you are experiencing same problem or not. Vilius Å umskas wrote: Sveiki, [...]]]>
Sveiki, Monday, November 16, 2009, 7:25:29 PM, you wrote: [...] It's always the same Segmentation Fault on line 554 at fsm.c. It is preceded by malloc errors, but not always crashes just right after them. For example: Nov 16 13:08:51 xserve3 com.apple.xsan[92] [...]]]>
What kind of errors are you getting in Xsan? Vilius Å umskas wrote: [...]]]>
[...] Sure :) I have bought two 4 port expansion licences to upgrade my switches from 16 to 20 ports. Vilius Å umskas LNK TV IT manager mob.: +370 614 75713 http://www.lnk.lt]]>
Hi, when we expanded customer's XSAN and needed to use more 5200s, we had to purchase not only XPACKs but also licensees to activate ISL ports to the standard 5200 config. It was 16port 5200 originally. Are you sure your ISLs are okay ? I mean , do you have enough licensses ? [...]]]>
I have Auto for IO Stream Guard and Enabled for Device Scan. Vilius Å umskas wrote: [...] Philadelphia, PA 19106 267-765-7000 (T) 267-765-7002 (F) www.cubistmediagroup.com]]>
I believe you need the 7.4.0.XX firmware. It is the firmware that no longer supports management with SanBox. You can only manage it with the Quicktool web GUI or the Enterprise Fabric Suite. My 10G are on default settings. -Phil What version exactly you are talking about? [...]]]>
What version exactly you are talking about? I'm running the newest firmware from 6.x series. Also do you know what settings must be on ISL ports? All other ports are configured according Apple docs, except for ISL as I don't know what settings they should be. [...]]]>
The 5200 has to have a firmware upgrade for compatibility with the 5600 and 10G. (It has to be one of the embedded web GUI versions). We have a similar setup working fine. Also check that all your ports are setup to the correct initiator settings. [...]]]>
Do you think I did right when I changed ISL this way: 1. Connected 10G cable. 2. Waited for the link activity led to show than ISL switched to that 10G. 3. Disconnected 2x2G ISL. Everything was done on running Xsan. Vilius Å umskas LNK TV IT manager mob.: +370 614 75713 http://www.lnk.lt [ [...]]]>
We still run a stack of six 5200s along with 9200 all connected with 10G Xpak cables. No problems here Thank you, Vitali Bystritski Sent from my iPhone On Nov 16, 2009, at 8:40 AM, Vilius Å umskas <email@hidden> wrote: [...]]]>
Hello list, Recently I have replaced two 2G interconnects with one 10G interconnect cable bought from Qlogic. Now all my volumes started to restart themselves once or two times a day. I have checked switches for errors on 10G ports, but everything seems good on that side. [...]]]>
Hi there, I'm using xsan2, and i think the followings messages are related to installed xsan. But in fact i don't know. Nov 16 13:00:03 slave sandboxd[2817]: sshd(2816) deny mach-per-user-lookup Nov 16 13:13:13: --- last message repeated 4 times --- Nov 16 13:13:13 slave servermgrd[58] [...]]]>
[...] I wanted to write "Copy hang" :) -- *Fabien COMBERNOUS* /unix system engineer/ www.kezia.com <http://www.kezia.com/> *Tel: +33 (0) 467 992 986* Kezia Group]]>
Hello there, After many tests, i changed SFP, now it is better. But i need to make some changes yet :) Some times during a copy from local hard drive of a client to xsan mounted volume, i have the following sequence : -1- starting sandboxd message about sshd. [...]]]>
Humm. there was definetly a metadata lun as per your config ... logically should a raid 1 of 2 drives ( 2 x 400G) . There was a generation of Xserveraid with 400gb drives. Any xserveraid lying somewhere ?? Putting metadata on something else is crazy... firewire or internal ... [...]]]>
Yes, it's possible that the configuration got scrambled, I have focused primarily on the cabling changes and issues associated with actual connectivity, as no one was in on the admin interface making changes prior to this problem occurring. Here is the . [...]]]>
Metadata can reside inside a data lun. It depends on the config. Maybe the missing lun pointing to something else/ Can you supply the xxxx.cfg file ? and the error in the system.log when trying to mount/assemble the volume ? serge On 12-Nov-09, at 3:10 PM, email@hidden wrote: [...]]]>
Agreed, such a non-standard setup would be quite unusual indeed. I'm really hung up on the fact that the metadata lun isn't on this RAID system. That's pretty hard to believe. Yet, clearly, here are all the volumes of the raid, with all the drives and space accounted for. [...]]]>
I haven't ever tried it, but in theory, I think so -- if there's no live failover between the two MDCs, a direct-connected drive (internal, firewire, possibly even iSCSI, anything that would show up in disk utility and would show up in 'cvlabel -l') can be used as a metadata drive. [...]]]>
Yes, the RAID systems checks out ok - and again, all the storage on the RAID system is accounted for and shows up in Xsan, disk utility, etc. Nothing appears to be missing from the raid. Here's the config files for the xsan. Attachment: conf.tar. [...]]]>
Good day, I'm working with an unfamiliar xsan system with a single volume ("media") that will not mount. It's looking for LUN1, which contains the meta data. The data LUNs show up fine, 4.77TB each x 4, but the 378GB metadata LUN1 does not show up in system profiler, in disk [...]]]>
[...] A replica for every host in Xsan network seems like too much overhead to me. Especially if there are more than 10. Eventually you will end up spending too much time debugging synchronization issues, etc. I'm running 3 Xserve's 1 OD master, 1 Replica/MDC backup, 1 MDC main for 18 node Xsan. [...]]]>
On Nov 12, 2009, at 9:17 AM, Fabien COMBERNOUS wrote: [...] I don't know if it is better to have only 2 OD (one master and one replica), or have each xsan host a local OD replica. And i foun how to modify the expected search policy. It's juste needed [...]]]>
On Nov 12, 2009, at 9:17 AM, Fabien COMBERNOUS wrote: Fabien COMBERNOUS wrote: [...] You probably don't want to do this. A list of replicas is kept on each machine connected to OD, and a replica will be automatically selected if the OD master is unavailable. [...]]]>
[...] I'm using xsan 2.2. In fact it appears that all hosts needs to have the same search policy. During the setup, i provided the address of my OD master to the wizard. And now all my hosts expects my OD master instead of the local replica. How to modify the expected search policy by xsan to point to the localhost replica ? -- *Fabien COMBERNOUS* /unix system engineer/ www.kezia.com <http://www.kezia.com/> *Tel: +33 (0) 467 992 986* Kezia Group]]>
Hi there, I would like to setup all my mdc as OD replica. This ensure continual availability of Open Directory service. But when i do this, in the Xsan Admin, my mdc appears with status "Incorrect search policy". Do you know how to solve this issue. [...]]]>
[...] I don't use a dedicated metadata network switch but a HP 8200zl. This switch is used only for my xsan lab. 2 vlan are configurated: one about public tcp/ip network, one about metadata tcp/ip network. I'll check connections. On all my xsan nodes i have this log in /var/log/system. [...]]]>
[...] Well your MDC simply looses the connection to the Volume. It could be metadata OR it could be FC network. As told earlier, Xsan is really complicated and you should check all the connections and settings. Even order of network connections on MDC and clients matter. [...]]]>
[...] I note your link and i'll use it. But i don't understand why you are thinking the problem is with fc network and not the metadata tcp/ip network. I'm using brocade b300 and xyratex storage. They are out of matrix. But unfortunately, i didn't choice the hardware. [...]]]>
Are your Fibre Channel switch and FC cards on MDC and clients configured correctly? What FC switch you have? Make sure it is configured according http://support.apple.com/kb/TA24251 Also what type of storage you are using? Xserve RAID? Promise? 3rd party? [...]]]>
[...] Ok. i configured dns. And i configured a new volume with .Trash file created by the touch command. But the system is still having some troubles. At the begining all was ok. But after a moment, i don't know why i have the following logs in syslog of the mdc : (On my lab 192.168.2. [...]]]>
public one the metadata dns works almost like a shadow government: some services need to resolve addresses on it just in case they are needed. On Nov 4, 2009, at 9:26 AM, Fabien COMBERNOUS wrote: Vitali Bystritski wrote: [...]]]>
[...] Lol :) Sorry for my bad english. -- *Fabien COMBERNOUS* /unix system engineer/ www.kezia.com <http://www.kezia.com/> *Tel: +33 (0) 467 992 986* Kezia Group]]>
[...] By "dustbin" you mean Trash folder? If yes, then take a look at the following documentation: http://support.apple.com/kb/TA24727?viewlocale=en_US]]>