I followed the instruction to setup the environment and run pretraining of the T5 baseline. I am using 4x A800 GPUs on 2 nodes so I tweaked the script a little bit. Here is the command I am using:
Here RANK the environment variable is 0 on the first node and 1 on the second node.
The pretraining seems to be stucked after the following logs:
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmprk24gayf, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpo1vx1gx5, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmprk24gayf, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpo1vx1gx5, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmprk24gayf, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpo1vx1gx5, retrying 198 more times
[2023-08-13 10:57:34,795][fairseq.data.iterators][INFO] - grouped total_num_itrs = 80180
[2023-08-13 10:57:34,799][fairseq.trainer][INFO] - begin training epoch 2
[2023-08-13 10:57:34,800][fairseq_cli.train][INFO] - Start iterating over samples
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.308136GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.308136GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpa8nzn28s, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpn36xafqb, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpa8nzn28s, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpn36xafqb, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpa8nzn28s, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp65tn9jz5, retrying 200 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.308136GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpn36xafqb, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.308136GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp65tn9jz5, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpsro1kc9c, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp65tn9jz5, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.308136GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpihqj3v92, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpsro1kc9c, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpihqj3v92, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpsro1kc9c, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.308136GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
[2023-08-13 10:57:35,455][numexpr.utils][INFO] - Note: detected 128 virtual cores but NumExpr set to maximum of 64, check "NUMEXPR_MAX_THREADS" environment variable.
[2023-08-13 10:57:35,456][numexpr.utils][INFO] - Note: NumExpr detected 128 cores but "NUMEXPR_MAX_THREADS" not set, so enforcing safe limit of 8.
[2023-08-13 10:57:35,456][numexpr.utils][INFO] - NumExpr defaulting to 8 threads.
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpihqj3v92, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.308136GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpwho59wvg, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp7i_f99vn, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpv48fb721, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpwho59wvg, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpg9n4lp3t, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp7i_f99vn, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp7stbg9e3, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpv48fb721, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpwho59wvg, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpg9n4lp3t, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp7i_f99vn, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp7stbg9e3, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpv48fb721, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp_xhq1bve, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpwho59wvg, retrying 197 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpg9n4lp3t, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp7i_f99vn, retrying 197 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp7stbg9e3, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpv48fb721, retrying 197 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp_xhq1bve, retrying 199 more times
../src/plasma/store.cc../src/plasma/store.cc../src/plasma/store.cc:::127412741274: : Allowing the Plasma store to use up to : Allowing the Plasma store to use up to Allowing the Plasma store to use up to 0.462204GB of memory.0.4622040.462204
GB of memory.GB of memory.../src/plasma/store.cc:
1297
../src/plasma/store.cc: :../src/plasma/store.ccStarting object store with directory 1297:/dev/shm: 1297 and huge page support Starting object store with directory : Starting object store with directory /dev/shm/dev/shmdisabled and huge page support and huge page support
disableddisabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.462204GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.308136GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp_xhq1bve, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.462204GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp27xexyt6, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpm3t5h5l6, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp27xexyt6, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpm3t5h5l6, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp27xexyt6, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.462204GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpm3t5h5l6, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.462204GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpt2f4086i, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpqrev4_d7, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp22mdeu4q, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpn8zcdjx3, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpt2f4086i, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpqrev4_d7, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpyp76x2qn, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp22mdeu4q, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpn8zcdjx3, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpt2f4086i, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpqrev4_d7, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpyp76x2qn, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp22mdeu4q, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpn8zcdjx3, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.594368GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.594368GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc../src/plasma/store.cc::12741274: : Allowing the Plasma store to use up to Allowing the Plasma store to use up to 0.594368GB of memory.0.594368
GB of memory.../src/plasma/store.cc:
1297../src/plasma/store.cc: :Starting object store with directory 1297/dev/shm: and huge page support Starting object store with directory /dev/shmdisabled and huge page support
disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpyp76x2qn, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.594368GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpiu3v5ik0, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpxbmo_moi, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpiu3v5ik0, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpxbmo_moi, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpiu3v5ik0, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.594368GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.594368GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpxbmo_moi, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp9f1r06z6, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpaadb9mhc, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp15em8bql, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmprnm3112l, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp3oaxfj3b, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp9f1r06z6, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpaadb9mhc, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp15em8bql, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmprnm3112l, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpzo2jf5x9, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp3oaxfj3b, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp9f1r06z6, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpaadb9mhc, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp15em8bql, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmprnm3112l, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpzo2jf5x9, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp3oaxfj3b, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.891552GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpaadb9mhc, retrying 197 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp15em8bql, retrying 197 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.891552GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmprnm3112l, retrying 197 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpzo2jf5x9, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpy97kbew0, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp3oaxfj3b, retrying 197 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.891552GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.462204GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.891552GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.891552GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpvaviv5wx, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpy97kbew0, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpvaviv5wx, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpy97kbew0, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.891552GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpvaviv5wx, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.891552GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp1tjq419a, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpuwdqy464, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp1tjq419a, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp7q7m4az9, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpjcy09frn, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp27_edadr, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmplv6l7nw8, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpuwdqy464, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp1tjq419a, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp7q7m4az9, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpjcy09frn, retrying 199 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.576898GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp27_edadr, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmplv6l7nw8, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpuwdqy464, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.594368GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp7q7m4az9, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.576898GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpjcy09frn, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.576898GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp27_edadr, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.576898GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmplv6l7nw8, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.576898GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmptnoleq2p, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpc2_y4hg5, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmptnoleq2p, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpc2_y4hg5, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmptnoleq2p, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.576898GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpsus12hjf, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpc2_y4hg5, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.576898GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpsus12hjf, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpv0uw6p8s, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpdhvfhjli, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpj6lwtwwh, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpsus12hjf, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp76p15ofh, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpv0uw6p8s, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpb87b9lza, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpdhvfhjli, retrying 199 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.865347GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpj6lwtwwh, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp76p15ofh, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpv0uw6p8s, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpb87b9lza, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpdhvfhjli, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpj6lwtwwh, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp76p15ofh, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpv0uw6p8s, retrying 197 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpb87b9lza, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.891552GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpdhvfhjli, retrying 197 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpj6lwtwwh, retrying 197 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp76p15ofh, retrying 197 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.865347GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.865347GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpb87b9lza, retrying 197 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpqqm4z_za, retrying 200 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.865347GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.865347GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp4_3ikzuy, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpqqm4z_za, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpjwwox8ok, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp4_3ikzuy, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpqqm4z_za, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpjwwox8ok, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp4_3ikzuy, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.865347GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.865347GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpjwwox8ok, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.206938GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpyjpk2a3a, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmps1d5gnw0, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmppcbelm9v, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmplyaw5xzp, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpxvx2xl5g, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpyjpk2a3a, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmps1d5gnw0, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmppcbelm9v, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmplyaw5xzp, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpxvx2xl5g, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpyjpk2a3a, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmps1d5gnw0, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmppcbelm9v, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmplyaw5xzp, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp_0bu499n, retrying 200 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.206938GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.206938GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpxvx2xl5g, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.206938GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.576898GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.206938GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpdunz7mkl, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp98v48lsg, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp_0bu499n, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpdunz7mkl, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp98v48lsg, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp_0bu499n, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpdunz7mkl, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp98v48lsg, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.310407GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc../src/plasma/store.cc::12741274: : Allowing the Plasma store to use up to Allowing the Plasma store to use up to 0.2069380.206938GB of memory.GB of memory.
../src/plasma/store.cc../src/plasma/store.cc::12971297: : Starting object store with directory Starting object store with directory /dev/shm/dev/shm and huge page support and huge page support disableddisabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp5j550t5e, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpvhj403ka, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpifz2dlz1, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp5j550t5e, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp42b62hz3, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp_3y02k40, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpvhj403ka, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpifz2dlz1, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp5j550t5e, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp42b62hz3, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp_3y02k40, retrying 199 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.310407GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpvhj403ka, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpdi3xuw9l, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpifz2dlz1, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.310407GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpbxlw3p_r, retrying 200 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.310407GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp42b62hz3, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.310407GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp_3y02k40, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpw6jqztgq, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpdi3xuw9l, retrying 199 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.865347GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpbxlw3p_r, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpw6jqztgq, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpdi3xuw9l, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpbxlw3p_r, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.502836GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.310407GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpw6jqztgq, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.310407GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpnktfo0dm, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpqon7mc6v, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpnktfo0dm, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpz9yvu6ak, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpqon7mc6v, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp3z517802, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpnktfo0dm, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpz9yvu6ak, retrying 199 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.502836GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpqon7mc6v, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.502836GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpmicmyzhq, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp3z517802, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpz9yvu6ak, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpsxflpclu, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpkvl78rmd, retrying 200 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.502836GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpmicmyzhq, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpo0dw02tb, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp3z517802, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpsxflpclu, retrying 199 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.502836GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpkvl78rmd, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpmicmyzhq, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.206938GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpo0dw02tb, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpsxflpclu, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpkvl78rmd, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.502836GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.754253GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpo0dw02tb, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.502836GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmplnlmm3io, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp5z7hg8y9, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp97yitntc, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmplnlmm3io, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp5z7hg8y9, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp97yitntc, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmplnlmm3io, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp5z7hg8y9, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpw30i3mb8, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp97yitntc, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmplnlmm3io, retrying 197 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp5z7hg8y9, retrying 197 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpw30i3mb8, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpzjsnhnsj, retrying 200 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.754253GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp639k100u, retrying 200 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.754253GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.754253GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpb54virj3, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpw30i3mb8, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpzjsnhnsj, retrying 199 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.310407GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp639k100u, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpb54virj3, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpzjsnhnsj, retrying 198 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmp639k100u, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.754253GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.754253GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmpb54virj3, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.754253GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmphq1j01h9, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmphq1j01h9, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmphq1j01h9, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.502836GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmprehp5o38, retrying 200 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmprehp5o38, retrying 199 more times
../src/plasma/io.cc:177: Connection to IPC socket failed for pathname /tmp/tmprehp5o38, retrying 198 more times
../src/plasma/store.cc:1274: Allowing the Plasma store to use up to 0.754253GB of memory.
../src/plasma/store.cc:1297: Starting object store with directory /dev/shm and huge page support disabled
I followed the instruction to setup the environment and run pretraining of the T5 baseline. I am using 4x A800 GPUs on 2 nodes so I tweaked the script a little bit. Here is the command I am using:
Here
RANK
the environment variable is0
on the first node and1
on the second node.The pretraining seems to be stucked after the following logs: