Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Alex wrote: > Hello experts, > > I read that raid software on linux is not cluster aware, so i'm trying to find > a solution to join together more computers to form a shared file system and > build a SAN (correctly if i am wrong), avoiding usage of raid software. > > Let say that I have: > - N computers (N>8) sharing their volumes (volX, where X=N). Each volX is > arround 120GB. > - M servers (M>3) - which are accessing a GFS volume (read/write) > - Other regular computers which are available if required. > > Now, I want: > - to build somehow a GFS on top of vol1, vol2, ... volN volumes with high data > availability and without a single point of failure. > If You want to use GFS, You would need one volX exported via iSCSI or GNDB to M servers. For HA you could use DRBD - two volumes vol1, vol2 created as /dev/drbd0 and exported to M servers. > - resulted logical volume to be used on SERVER1, SERVER2 and SERVER3 > (read/write access) > With CLVM You could have logical volumes for GFS fs. > So, how can i do that? Is that possible? > > Can somebody suggest me an scenario regarding how to grup machines? > > My scenario (before to start): > > Step1. Grup computerX together, 2 by 2 using drbd and create network raid 1 > mirrors for each, in order to produce: > > On computer1 and computer2: > - /dev/drbd0 (120GB size, contain vol1 <-> vol2 mirrored) > On computer3 and computer4: > - /dev/drbd1 (120GB size, contain vol3 <-> vol4 mirrored) > On computer5 and computer6: > - /dev/drbd2 (120GB size, contain vol5 <-> vol6 mirrored) > On computer7 and computer8: > - /dev/drbd3 (120GB size, contain vol7 <-> vol8 mirrored) > > Does it mean that will result 4 CLUSTERS (4 different cluster.conf files). Is > that correct? If yes, how can i join together resulted volumes /dev/drbd*? > > Will be ok to export resulted /dev/drbd* volumes using ataoe (iscsi or gnbd) > to our SERVERS and after that, on SERVER1 for example, to manipulate and join > imported volumes (/dev/import0, /dev/import1, /dev/import2, /dev/import3) > using clvm like on the next step below? > > Step2. On SERVER1, join resulted volumes together using lvm and create a > logical volume (480GB) > pvcreate /dev/import0 /dev/import1 /dev/import2 /dev/import3 > vgcreate myvg ... > lvcreate mylvm ... > > Now, mylvm is grouping import0 up to import4 in one logical volume which is in > fact drbd0+drbd1+drbd2+drbd3. > > What will be if at this stage, will not be available any fencing? I know that > GFS require fencing in all circumstances. > > How can be implemented fencing? ATAoE and ISCSI does not provide any builtin > fencing mechanism, so, maybe GNBD sould be used to export /dev/drbd* (because > it has fencing builtin). Is that correct? Is possible? If, yes, should i > create only one cluster grouping: > - all N computers together > or > - all N computers plus our 3 SERVERS together? > > Step3. format /dev/myvg/mylv using GFS: > mkfs.gfs -p lock_dlm -t cluster:data -j 3 /dev/myvg/mylv > > Step4. Create other cluster containing only 3 nodes (SERVER1 up to SERVER3) > which will mount and use resulted shared GFS volume: > > mount /dev/myvg/mylv /var/www/data on all our servers. > > Any ideas? > Your solution isn't good for GFS fs. You can also use other HA/Cluster fs like: hadoop, gluster, kosmos-fs, mogile and much more. Best regards Maciej Bogucki